Re: Hipersockets and Linux...
Yes, that did the trick. Thanks for the solution with such details! Frank M. Ramaekers Jr. -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Vic Cross Sent: Monday, March 8, 2021 9:43 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets and Linux... External Message – Think Before You Click Frank wrote: >Okay, following along in "IBM HiperSockets Implementation Guide" >Chapter 3 "Software configurations for HiperSockets. The command >"ifconfig enccw0.0.0800 192.168.250.88 netmask 255.255.255.0 up" >seems to work, only temporarily. > ># ping 192.168.250.101 >PING 192.168.250.101 (192.168.250.101) 56(84) bytes of data. >64 bytes from 192.168.250.101: icmp_seq=1 ttl=255 time=0.362 ms >64 bytes from 192.168.250.101: icmp_seq=2 ttl=255 time=0.313 ms > >Then just a few seconds later: ># ping 192.168.250.101 >PING 192.168.250.101 (192.168.250.101) 56(84) bytes of data. >From 32.140.72.137 icmp_seq=9 Packet filtered From 32.140.72.137 >icmp_seq=19 Packet filtered From 32.140.72.137 icmp_seq=20 Packet >filtered Your ifconfig has proven that the interface works, which is great, but using ifconfig (or any of the manual interface config commands, such as ip) is a very temporary config method. At best it lasts until the next reboot, but these days often not even that long. I suspect that NetworkManager is trying to configure the interface for you. Its standard treatment of unconfigured interfaces is to use DHCP to get an address, so it's probably the DHCP client that is clearing what you do with ifconfig. You should have an ifcfg-* file in /etc/sysconfig/network-scripts/ for your OSA/VSwitch interface... Copy this file to ifcfg-enccw0.0.0800 and make all the required changes inside there (you can delete the UUID line, if you still have an OSA/VSwitch to get to most things then delete the GATEWAY line, and if you want to talk to z/OS make sure that in the OPTIONS line you have "layer2=0"). Then restart NetworkManager (systemctl restart NetworkManager.service). NM will then see your HiperSockets as another "System" connection and manage it for you. Regards, Vic -- Vic Cross Solutions Engineer, Z Acceleration Team IBM Z (Worldwide) E-mail: viccr...@au1.ibm.com Twitter: @viccross -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit https://urldefense.com/v3/__http://www2.marist.edu/htbin/wlvindex?LINUX-390__;!!GePVz-XrgrDUrw!-p10A02CI_MFnOGPTEJqP4hcHYXgnF5fGmih5a_R0dCyG-ktv18zsEdEVTDe4UaZ$ -- This message contains information which is privileged and confidential and is solely for the use of the intended recipient. If you are not the intended recipient, be aware that any review, disclosure, copying, distribution, or use of the contents of this message is strictly prohibited. If you have received this in error, please destroy it immediately and notify us at privacy...@globe.life. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www2.marist.edu/htbin/wlvindex?LINUX-390
Re: Hipersockets and Linux...
Frank wrote: >Okay, following along in "IBM HiperSockets Implementation Guide" >Chapter 3 "Software configurations for HiperSockets. The command >"ifconfig enccw0.0.0800 192.168.250.88 netmask 255.255.255.0 up" >seems to work, only temporarily. > ># ping 192.168.250.101 >PING 192.168.250.101 (192.168.250.101) 56(84) bytes of data. >64 bytes from 192.168.250.101: icmp_seq=1 ttl=255 time=0.362 ms >64 bytes from 192.168.250.101: icmp_seq=2 ttl=255 time=0.313 ms > >Then just a few seconds later: ># ping 192.168.250.101 >PING 192.168.250.101 (192.168.250.101) 56(84) bytes of data. >From 32.140.72.137 icmp_seq=9 Packet filtered >From 32.140.72.137 icmp_seq=19 Packet filtered >From 32.140.72.137 icmp_seq=20 Packet filtered Your ifconfig has proven that the interface works, which is great, but using ifconfig (or any of the manual interface config commands, such as ip) is a very temporary config method. At best it lasts until the next reboot, but these days often not even that long. I suspect that NetworkManager is trying to configure the interface for you. Its standard treatment of unconfigured interfaces is to use DHCP to get an address, so it's probably the DHCP client that is clearing what you do with ifconfig. You should have an ifcfg-* file in /etc/sysconfig/network-scripts/ for your OSA/VSwitch interface... Copy this file to ifcfg-enccw0.0.0800 and make all the required changes inside there (you can delete the UUID line, if you still have an OSA/VSwitch to get to most things then delete the GATEWAY line, and if you want to talk to z/OS make sure that in the OPTIONS line you have "layer2=0"). Then restart NetworkManager (systemctl restart NetworkManager.service). NM will then see your HiperSockets as another "System" connection and manage it for you. Regards, Vic -- Vic Cross Solutions Engineer, Z Acceleration Team IBM Z (Worldwide) E-mail: viccr...@au1.ibm.com Twitter: @viccross -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www2.marist.edu/htbin/wlvindex?LINUX-390
Re: hipersockets
On Friday, 11/20/2015 at 02:28 GMT, Grzegorz Powiedziuk wrote: > ohhh I see what you are saying (I think so). > > I was thinking about standard usage of vswitch bridge so a scenario where > all guests have real dedicated hipersockets and vswitch has one of these > hipersockets as well. This gives hipersocket virtual machines a way to talk > to each other (via hipersocket network and shared chpid) plus it gives them > access to oustide network via vswitch. > > But you are saying (I think) that I could have ONLY vswitch (on all lpars) > using the hipersocket bridge port and all the guest have connected to the > same vswitch (layer two so oracle should be happy) using virtual qdio nic > cards? That sounds very cool. I don't have RACF yet but it will give me a > standard vswitch access list control of vlans which is something. > > In scenario like this, I guess I don't need to specify uplink device for > the vswitch so I could keep it all inside of a CEC. > > I wonder how much how much would it impact performance. Sorry if I've mislead you. The bridge can't itself enforce VLAN assignments since the guests use real HiperSockets, not virtual ones. People who want more VLAN controls in (real) OSA and (real) HiperSockets need to make their wants known via RFEs. Alan Altmark Senior Managing z/VM and Linux Consultant Lab Services System z Delivery Practice IBM Systems & Technology Group ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets
On Friday, 11/20/2015 at 02:02 GMT, David Kreuter wrote: > Another advantage is that it uses only one hipersocket triplet whereas > dedicated hipersocket will end up using one triplet per Linux virtual. > You can even use the same hiper triplet on each LPAR. > There is a limit to how many LPARs can connect to the hiper. I've been > scalded by this a few times. David, each guest uses a HiperSocket triplet, as does the bridge itself. There is no (effective) limit on the number of guests that can use the bridge. A maximum of 5 *bridges* can connect to the chpid, but you can have as many LPARs as you like. (Only one of the bridges is active at any one time.) Alan Altmark Senior Managing z/VM and Linux Consultant Lab Services System z Delivery Practice IBM Systems & Technology Group ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets
ohhh I see what you are saying (I think so). I was thinking about standard usage of vswitch bridge so a scenario where all guests have real dedicated hipersockets and vswitch has one of these hipersockets as well. This gives hipersocket virtual machines a way to talk to each other (via hipersocket network and shared chpid) plus it gives them access to oustide network via vswitch. But you are saying (I think) that I could have ONLY vswitch (on all lpars) using the hipersocket bridge port and all the guest have connected to the same vswitch (layer two so oracle should be happy) using virtual qdio nic cards? That sounds very cool. I don't have RACF yet but it will give me a standard vswitch access list control of vlans which is something. In scenario like this, I guess I don't need to specify uplink device for the vswitch so I could keep it all inside of a CEC. I wonder how much how much would it impact performance. Thanks! Gregory 2015-11-20 9:00 GMT-05:00 David Kreuter : > Gregory: It does help. As a VSWITCH it can be protected by RACF hence > the VLANs can be RACF protected too. > Another advantage is that it uses only one hipersocket triplet whereas > dedicated hipersocket will end up using one triplet per Linux virtual. > You can even use the same hiper triplet on each LPAR. > There is a limit to how many LPARs can connect to the hiper. I've been > scalded by this a few times. > David > > > ---- Original Message > Subject: Re: hipersockets > From: Grzegorz Powiedziuk > Date: Fri, November 20, 2015 8:47 am > To: LINUX-390@VM.MARIST.EDU > > Thanks Alan. > HiperSocket VSWITCH Bridge will not help when it comes to isolation, > right? Vswitch simply acts as a bridge for hipersocket network using > using > one of the real hipersocket devices as one of its own interfaces (bridge > port). Via this bridge Hipersocket network gets access to external > network > but doesn't give more control on who can talk to who inside of CEC, does > it? > thanks > Gregory > > > > 2015-11-20 0:47 GMT-05:00 Alan Altmark : > > > On Thursday, 11/19/2015 at 08:35 GMT, Grzegorz Powiedziuk > > wrote: > > > I thought about doing vswitch but then AFIK I would end up with with > > > virtual hipersockets on linux guest. > > > > Linux guests can use real HiperSockets with the HiperSocket VSWITCH > bridge > > on z/VM. Their traffic will automatically be bridged to a physical LAN > > that can be accessed by z/OS. z/OS doesn't support the HiperSocket > > technology that would let it participate in a direct HiperSocket > > connection with the Linux guests on the bridge. > > > > Alan Altmark > > > > Senior Managing z/VM and Linux Consultant > > Lab Services System z Delivery Practice > > IBM Systems & Technology Group > > ibm.com/systems/services/labservices > > office: 607.429.3323 > > mobile; 607.321.7556 > > alan_altm...@us.ibm.com > > IBM Endicott > > > > -- > > For LINUX-390 subscribe / signoff / archive access instructions, > > send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or > > visit > > http://www.marist.edu/htbin/wlvindex?LINUX-390 > > -- > > For more information on Linux on System z, visit > > http://wiki.linuxvm.org/ > > > > -- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or > visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > -- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ > > -- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or > visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > -- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ > -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets
Gregory: It does help. As a VSWITCH it can be protected by RACF hence the VLANs can be RACF protected too. Another advantage is that it uses only one hipersocket triplet whereas dedicated hipersocket will end up using one triplet per Linux virtual. You can even use the same hiper triplet on each LPAR. There is a limit to how many LPARs can connect to the hiper. I've been scalded by this a few times. David Original Message Subject: Re: hipersockets From: Grzegorz Powiedziuk Date: Fri, November 20, 2015 8:47 am To: LINUX-390@VM.MARIST.EDU Thanks Alan. HiperSocket VSWITCH Bridge will not help when it comes to isolation, right? Vswitch simply acts as a bridge for hipersocket network using using one of the real hipersocket devices as one of its own interfaces (bridge port). Via this bridge Hipersocket network gets access to external network but doesn't give more control on who can talk to who inside of CEC, does it? thanks Gregory 2015-11-20 0:47 GMT-05:00 Alan Altmark : > On Thursday, 11/19/2015 at 08:35 GMT, Grzegorz Powiedziuk > wrote: > > I thought about doing vswitch but then AFIK I would end up with with > > virtual hipersockets on linux guest. > > Linux guests can use real HiperSockets with the HiperSocket VSWITCH bridge > on z/VM. Their traffic will automatically be bridged to a physical LAN > that can be accessed by z/OS. z/OS doesn't support the HiperSocket > technology that would let it participate in a direct HiperSocket > connection with the Linux guests on the bridge. > > Alan Altmark > > Senior Managing z/VM and Linux Consultant > Lab Services System z Delivery Practice > IBM Systems & Technology Group > ibm.com/systems/services/labservices > office: 607.429.3323 > mobile; 607.321.7556 > alan_altm...@us.ibm.com > IBM Endicott > > -- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or > visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > -- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ > -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets
Thanks Alan. HiperSocket VSWITCH Bridge will not help when it comes to isolation, right? Vswitch simply acts as a bridge for hipersocket network using using one of the real hipersocket devices as one of its own interfaces (bridge port). Via this bridge Hipersocket network gets access to external network but doesn't give more control on who can talk to who inside of CEC, does it? thanks Gregory 2015-11-20 0:47 GMT-05:00 Alan Altmark : > On Thursday, 11/19/2015 at 08:35 GMT, Grzegorz Powiedziuk > wrote: > > I thought about doing vswitch but then AFIK I would end up with with > > virtual hipersockets on linux guest. > > Linux guests can use real HiperSockets with the HiperSocket VSWITCH bridge > on z/VM. Their traffic will automatically be bridged to a physical LAN > that can be accessed by z/OS. z/OS doesn't support the HiperSocket > technology that would let it participate in a direct HiperSocket > connection with the Linux guests on the bridge. > > Alan Altmark > > Senior Managing z/VM and Linux Consultant > Lab Services System z Delivery Practice > IBM Systems & Technology Group > ibm.com/systems/services/labservices > office: 607.429.3323 > mobile; 607.321.7556 > alan_altm...@us.ibm.com > IBM Endicott > > -- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or > visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > -- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ > -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets
On Thursday, 11/19/2015 at 08:35 GMT, Grzegorz Powiedziuk wrote: > I thought about doing vswitch but then AFIK I would end up with with > virtual hipersockets on linux guest. Linux guests can use real HiperSockets with the HiperSocket VSWITCH bridge on z/VM. Their traffic will automatically be bridged to a physical LAN that can be accessed by z/OS. z/OS doesn't support the HiperSocket technology that would let it participate in a direct HiperSocket connection with the Linux guests on the bridge. Alan Altmark Senior Managing z/VM and Linux Consultant Lab Services System z Delivery Practice IBM Systems & Technology Group ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets
Thanks David. I thought about doing vswitch but then AFIK I would end up with with virtual hipersockets on linux guest. And I've read in IBMs redbook for oracle 12: IBM HiperSockets™ are certified and supported for the private network. Only a network that is configured with *real* HiperSockets is possible, as z/VM guest LAN HiperSockets cannot be configured on layer 2, which is required for ARP. Gregory 2015-11-19 15:20 GMT-05:00 David Kreuter : > Hi - I've done the hipersocket VLAN implementation. It works well and of > course Alan's comments are correct. > > Another approach I've used is to create a VSWITCH on each LPAR using the > same set of OSAs. Now when you use VLANs on this VSWITCH RACF can be > involved for better protection. > > OK won't be as fast as hipersocket but it doesn't go far out of the box > either. > David Kreuter > > > > Original Message > Subject: Re: hipersockets > From: Alan Altmark > Date: Thu, November 19, 2015 3:05 pm > To: LINUX-390@VM.MARIST.EDU > > On Thursday, 11/19/2015 at 07:38 GMT, Grzegorz Powiedziuk > wrote: > > From what I've learned so far, In order to achieve this, we need to have > a > > shared chpid between LPARS. Hipersockets on the same chpid can > communicate > > with each other. > > Hosts using the same VLAN on the same HiperSocket chpid can talk to each > > other. There are no controls on the VLAN ID that a host is permitted to > use, so from a security perspective, don't rely on HiperSocket VLAN > controls. > > > Ok, we've done that. We have defined a set of hipersockets on one chipd > for > > every LPAR and it works. Linux in one LPAR can talk to another linux in > > different lpar. > : > > Do I need to have a separate chpid for every cluster? Doesn't really > make > > sense, does it? > > Am I missing something? > > It depends entirely on your security posture. If you need enforced > isolation of each pair, then you need one chpid per pair. > > Alan Altmark > > Senior Managing z/VM and Linux Consultant > Lab Services System z Delivery Practice > IBM Systems & Technology Group > ibm.com/systems/services/labservices > office: 607.429.3323 > mobile; 607.321.7556 > alan_altm...@us.ibm.com > IBM Endicott > > -- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or > visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > -- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ > > -- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or > visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > -- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ > -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets
Thank you for a really quick answer Alan. So I did get it right more or less. I didn't know that I can do vlans which will make things cleaner to some extent. But I was hopping for a different answer when it comes to security. We will have at least non-prod and prod environments on separate chpids then. Thank you Gregory 2015-11-19 15:05 GMT-05:00 Alan Altmark : > On Thursday, 11/19/2015 at 07:38 GMT, Grzegorz Powiedziuk > wrote: > > From what I've learned so far, In order to achieve this, we need to have > a > > shared chpid between LPARS. Hipersockets on the same chpid can > communicate > > with each other. > > Hosts using the same VLAN on the same HiperSocket chpid can talk to each > other. There are no controls on the VLAN ID that a host is permitted to > use, so from a security perspective, don't rely on HiperSocket VLAN > controls. > > > Ok, we've done that. We have defined a set of hipersockets on one chipd > for > > every LPAR and it works. Linux in one LPAR can talk to another linux in > > different lpar. > : > > Do I need to have a separate chpid for every cluster? Doesn't really > make > > sense, does it? > > Am I missing something? > > It depends entirely on your security posture. If you need enforced > isolation of each pair, then you need one chpid per pair. > > Alan Altmark > > Senior Managing z/VM and Linux Consultant > Lab Services System z Delivery Practice > IBM Systems & Technology Group > ibm.com/systems/services/labservices > office: 607.429.3323 > mobile; 607.321.7556 > alan_altm...@us.ibm.com > IBM Endicott > > -- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or > visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > -- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ > -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets
Hi - I've done the hipersocket VLAN implementation. It works well and of course Alan's comments are correct. Another approach I've used is to create a VSWITCH on each LPAR using the same set of OSAs. Now when you use VLANs on this VSWITCH RACF can be involved for better protection. OK won't be as fast as hipersocket but it doesn't go far out of the box either. David Kreuter Original Message Subject: Re: hipersockets From: Alan Altmark Date: Thu, November 19, 2015 3:05 pm To: LINUX-390@VM.MARIST.EDU On Thursday, 11/19/2015 at 07:38 GMT, Grzegorz Powiedziuk wrote: > From what I've learned so far, In order to achieve this, we need to have a > shared chpid between LPARS. Hipersockets on the same chpid can communicate > with each other. Hosts using the same VLAN on the same HiperSocket chpid can talk to each other. There are no controls on the VLAN ID that a host is permitted to use, so from a security perspective, don't rely on HiperSocket VLAN controls. > Ok, we've done that. We have defined a set of hipersockets on one chipd for > every LPAR and it works. Linux in one LPAR can talk to another linux in > different lpar. : > Do I need to have a separate chpid for every cluster? Doesn't really make > sense, does it? > Am I missing something? It depends entirely on your security posture. If you need enforced isolation of each pair, then you need one chpid per pair. Alan Altmark Senior Managing z/VM and Linux Consultant Lab Services System z Delivery Practice IBM Systems & Technology Group ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets
On Thursday, 11/19/2015 at 07:38 GMT, Grzegorz Powiedziuk wrote: > From what I've learned so far, In order to achieve this, we need to have a > shared chpid between LPARS. Hipersockets on the same chpid can communicate > with each other. Hosts using the same VLAN on the same HiperSocket chpid can talk to each other. There are no controls on the VLAN ID that a host is permitted to use, so from a security perspective, don't rely on HiperSocket VLAN controls. > Ok, we've done that. We have defined a set of hipersockets on one chipd for > every LPAR and it works. Linux in one LPAR can talk to another linux in > different lpar. : > Do I need to have a separate chpid for every cluster? Doesn't really make > sense, does it? > Am I missing something? It depends entirely on your security posture. If you need enforced isolation of each pair, then you need one chpid per pair. Alan Altmark Senior Managing z/VM and Linux Consultant Lab Services System z Delivery Practice IBM Systems & Technology Group ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets Early Completion Queue Support
On Thu, 2014-05-29 at 17:26 +, Gerard Howells wrote: > Hello All, > > As a result of some issues we've been seeing we're looking to implement the > z/VM Hipersockets Early Completion queue feature in our systems. We run VM > 6.2 with z/VSE 5.1 and a mixture of SLES 10 SP4 and 11 SP3 guests. I know > that it's supported in SLES 11 my question is whether it's supported in SLES > 10? > > Thanks in advance! > > Gerard Howells Gerard, currently Linux exploits HiperSockets Completion Queues only for the address family AF_IUCV (in case of HiperSockets transport), i.e. when running an AF_IUCV socket program. It is not used for standard IP-traffic across HiperSockets. Regards, Ursula Braun, IBM Germany, Linux on System z development -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets Early Completion Queue Support
>>> On 5/29/2014 at 01:26 PM, Gerard Howells wrote: > As a result of some issues we've been seeing we're looking to implement the > z/VM Hipersockets Early Completion queue feature in our systems. We run VM > 6.2 with z/VSE 5.1 and a mixture of SLES 10 SP4 and 11 SP3 guests. I know > that it's supported in SLES 11 my question is whether it's supported in SLES > 10? Where did you see a reference to that for SLES11? I can't find any (with a few quick searches). All I can find is z/VM references. Mark Post -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets between RHEL and z/OS
Thanks for helping me work through this... There needs to be a new redbook, titled: "The little things that you forget along the way". Thanks again, Offer. David Diep What would you like to see in the renovated Martin Luther King Jr. Memorial Library? Share your thoughts on the library's spaces and services at 'DC Public Librarys <http://www.dclibrary.org/> online idea community at dclibrary.ideascale.com <http://www.dclibrary.ideascale.com> -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Offer Baruch Sent: Monday, November 04, 2013 12:54 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets between RHEL and z/OS First of all i am glad to here it works... The home statment order only matters with vipa addresses... The way it works is all addresses after the vipa are sourced by it if the sourcevipa parameter is specified... If you want somthing to be excluded from that vipa you must place it before the vipa statment... It is all in the books :-) A few years back it was strange for me to read it as well... Offer Baruch On Nov 4, 2013 7:41 PM, "Diep, David (OCTO-Contractor)" wrote: > Offer, > > It does work... why on earth? > > I see from a packet trace that anything going to a Hipersockets > address > (5.5.5.0) has a source IP address of 5.5.5.x, while anything going > out normal OSAD, leaves with the VIPA address. > > What is with the HOME statement? I recall from OS390 TCPIP that the > order made a difference, but I don't recall why? > > David Diep > > > > What would you like to see in the renovated Martin Luther King Jr. > Memorial Library? > Share your thoughts on the library's spaces and services at 'DC Public > Librarys <http://www.dclibrary.org/> online idea community at > dclibrary.ideascale.com <http://www.dclibrary.ideascale.com> > > > > > > -Original Message- > From: Diep, David (OCTO-Contractor) > Sent: Monday, November 04, 2013 12:02 PM > To: 'Linux on 390 Port' > Subject: RE: Hipersockets between RHEL and z/OS > > EZZ2350I MVS TCP/IP NETSTAT CS V1R13 TCPIP Name: TCPIP > EZZ2700I Home address list: > EZZ2701I Address Link Flg > EZZ2703I 10.27.42.1 VLNKOS42 P > EZZ2703I 10.82.10.13 LOSA42P0 > EZZ2703I 10.82.10.81 LOSA43P0 > EZZ2703I 5.5.5.42 HSLINK42 > EZZ2703I 127.0.0.1LOOPBACK > > -Original Message- > From: Diep, David (OCTO-Contractor) > Sent: Monday, November 04, 2013 12:01 PM > To: LINUX-390@VM.MARIST.EDU > Subject: RE: Hipersockets between RHEL and z/OS > > Offer, > > I have a test system... let me give it a shot. > > Right now, I know for a fact that my HS Home IP address is the last > one listed. > > Thanks, > > David Diep > > > -Original Message- > From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of > Offer Baruch > Sent: Monday, November 04, 2013 11:48 AM > To: LINUX-390@VM.MARIST.EDU > Subject: Re: Hipersockets between RHEL and z/OS > > Hi > > You might be correct and this is a sourcevipa issue but i cant tell as > i am lacking some data... > Not many know this but the order of the home statment matters... you > must place your hiper sockets interface before your vipa interface if > you dont want that vipa to be applied to that interface. > If you send the output of the following tso commands we can know for sure: > Tso netstat home > Tso netstat route > > That should give us the answer i think... > > Offer > On Nov 4, 2013 3:11 PM, "Diep, David (OCTO-Contractor)" > > > wrote: > > > Hi Offer, > > > > All IPtables are off, both for v4 and v6. > > > > Pinging from z/OS to zLinux does not work, nor does traceroute yield > > anything. With all this said, I think it has something to do with > > the way I have routing set up on the z/OS side. For all OSA traffic, > > I use OSPF while for Hipersockets, I use static routing. > > > > - The reason I think this is because when I do a traceroute from > > z/OS to z/VM, it looks like the route is by way of my VLAN through > > the OSAs, though I have defined it for all traffic to 5.5.5.0/24 to > > be routed through the Hipersockets link: > > > > CS V1R13: Traceroute to 5.5.5.2 (5.5.5.2): > > 1 10.82.10.44 (10.82.10.44) 6 ms 10.82.10.121 (10.82.10.121) 0 ms > > 10.82.10.44 > > (10.82.10.44) 0 ms > > *** > > > > Just some notes... I do not have an OSPF interface defined for > > Hipersockets. When I cancel the OMPROUTE started task, I am unable > > to ping or reach other z/OS hiper
Re: Hipersockets between RHEL and z/OS
First of all i am glad to here it works... The home statment order only matters with vipa addresses... The way it works is all addresses after the vipa are sourced by it if the sourcevipa parameter is specified... If you want somthing to be excluded from that vipa you must place it before the vipa statment... It is all in the books :-) A few years back it was strange for me to read it as well... Offer Baruch On Nov 4, 2013 7:41 PM, "Diep, David (OCTO-Contractor)" wrote: > Offer, > > It does work... why on earth? > > I see from a packet trace that anything going to a Hipersockets address > (5.5.5.0) has a source IP address of 5.5.5.x, while anything going out > normal OSAD, leaves with the VIPA address. > > What is with the HOME statement? I recall from OS390 TCPIP that the order > made a difference, but I don't recall why? > > David Diep > > > > What would you like to see in the renovated Martin Luther King Jr. > Memorial Library? > Share your thoughts on the library's spaces and services at 'DC Public > Librarys <http://www.dclibrary.org/> online idea community at > dclibrary.ideascale.com <http://www.dclibrary.ideascale.com> > > > > > > -Original Message- > From: Diep, David (OCTO-Contractor) > Sent: Monday, November 04, 2013 12:02 PM > To: 'Linux on 390 Port' > Subject: RE: Hipersockets between RHEL and z/OS > > EZZ2350I MVS TCP/IP NETSTAT CS V1R13 TCPIP Name: TCPIP > EZZ2700I Home address list: > EZZ2701I Address Link Flg > EZZ2703I 10.27.42.1 VLNKOS42 P > EZZ2703I 10.82.10.13 LOSA42P0 > EZZ2703I 10.82.10.81 LOSA43P0 > EZZ2703I 5.5.5.42 HSLINK42 > EZZ2703I 127.0.0.1 LOOPBACK > > -Original Message- > From: Diep, David (OCTO-Contractor) > Sent: Monday, November 04, 2013 12:01 PM > To: LINUX-390@VM.MARIST.EDU > Subject: RE: Hipersockets between RHEL and z/OS > > Offer, > > I have a test system... let me give it a shot. > > Right now, I know for a fact that my HS Home IP address is the last one > listed. > > Thanks, > > David Diep > > > -Original Message- > From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of > Offer Baruch > Sent: Monday, November 04, 2013 11:48 AM > To: LINUX-390@VM.MARIST.EDU > Subject: Re: Hipersockets between RHEL and z/OS > > Hi > > You might be correct and this is a sourcevipa issue but i cant tell as i > am lacking some data... > Not many know this but the order of the home statment matters... you must > place your hiper sockets interface before your vipa interface if you dont > want that vipa to be applied to that interface. > If you send the output of the following tso commands we can know for sure: > Tso netstat home > Tso netstat route > > That should give us the answer i think... > > Offer > On Nov 4, 2013 3:11 PM, "Diep, David (OCTO-Contractor)" > > wrote: > > > Hi Offer, > > > > All IPtables are off, both for v4 and v6. > > > > Pinging from z/OS to zLinux does not work, nor does traceroute yield > > anything. With all this said, I think it has something to do with the > > way I have routing set up on the z/OS side. For all OSA traffic, I use > > OSPF while for Hipersockets, I use static routing. > > > > - The reason I think this is because when I do a traceroute from z/OS > > to z/VM, it looks like the route is by way of my VLAN through the > > OSAs, though I have defined it for all traffic to 5.5.5.0/24 to be > > routed through the Hipersockets link: > > > > CS V1R13: Traceroute to 5.5.5.2 (5.5.5.2): > > 1 10.82.10.44 (10.82.10.44) 6 ms 10.82.10.121 (10.82.10.121) 0 ms > > 10.82.10.44 > > (10.82.10.44) 0 ms > > *** > > > > Just some notes... I do not have an OSPF interface defined for > > Hipersockets. When I cancel the OMPROUTE started task, I am unable to > > ping or reach other z/OS hipersocket addresses. I do have sourceVIPA > > turned on... would this be the issue? At this point, I'm not too sure > > why its not routing via the static route. > > > > IPCONFIG IGNOREREDIRECT > > NOSYSPLEXROUTING > > MULTIPATH > > DATAGRAMFWD > > SOURCEVIPA > > PATHMTUDISC > > IPSECURITY > > ; > > Beginroutes > > route 5.5.5.0 255.255.255.0 = HSLINK28 MTU 57344 NODELAYA > > Endroutes > > ; > > PRIMARYINTERFACE VLNKOS28 > > > > > > Omproute options: > > > > AS_BOUNDARY_ROUTING > > Import_RIP_Routes=No > > Import_Static_Rou
Re: Hipersockets between RHEL and z/OS
Offer, It does work... why on earth? I see from a packet trace that anything going to a Hipersockets address (5.5.5.0) has a source IP address of 5.5.5.x, while anything going out normal OSAD, leaves with the VIPA address. What is with the HOME statement? I recall from OS390 TCPIP that the order made a difference, but I don't recall why? David Diep What would you like to see in the renovated Martin Luther King Jr. Memorial Library? Share your thoughts on the library's spaces and services at 'DC Public Librarys <http://www.dclibrary.org/> online idea community at dclibrary.ideascale.com <http://www.dclibrary.ideascale.com> -Original Message- From: Diep, David (OCTO-Contractor) Sent: Monday, November 04, 2013 12:02 PM To: 'Linux on 390 Port' Subject: RE: Hipersockets between RHEL and z/OS EZZ2350I MVS TCP/IP NETSTAT CS V1R13 TCPIP Name: TCPIP EZZ2700I Home address list: EZZ2701I Address Link Flg EZZ2703I 10.27.42.1 VLNKOS42 P EZZ2703I 10.82.10.13 LOSA42P0 EZZ2703I 10.82.10.81 LOSA43P0 EZZ2703I 5.5.5.42 HSLINK42 EZZ2703I 127.0.0.1LOOPBACK -Original Message- From: Diep, David (OCTO-Contractor) Sent: Monday, November 04, 2013 12:01 PM To: LINUX-390@VM.MARIST.EDU Subject: RE: Hipersockets between RHEL and z/OS Offer, I have a test system... let me give it a shot. Right now, I know for a fact that my HS Home IP address is the last one listed. Thanks, David Diep -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Offer Baruch Sent: Monday, November 04, 2013 11:48 AM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets between RHEL and z/OS Hi You might be correct and this is a sourcevipa issue but i cant tell as i am lacking some data... Not many know this but the order of the home statment matters... you must place your hiper sockets interface before your vipa interface if you dont want that vipa to be applied to that interface. If you send the output of the following tso commands we can know for sure: Tso netstat home Tso netstat route That should give us the answer i think... Offer On Nov 4, 2013 3:11 PM, "Diep, David (OCTO-Contractor)" wrote: > Hi Offer, > > All IPtables are off, both for v4 and v6. > > Pinging from z/OS to zLinux does not work, nor does traceroute yield > anything. With all this said, I think it has something to do with the > way I have routing set up on the z/OS side. For all OSA traffic, I use > OSPF while for Hipersockets, I use static routing. > > - The reason I think this is because when I do a traceroute from z/OS > to z/VM, it looks like the route is by way of my VLAN through the > OSAs, though I have defined it for all traffic to 5.5.5.0/24 to be > routed through the Hipersockets link: > > CS V1R13: Traceroute to 5.5.5.2 (5.5.5.2): > 1 10.82.10.44 (10.82.10.44) 6 ms 10.82.10.121 (10.82.10.121) 0 ms > 10.82.10.44 > (10.82.10.44) 0 ms > *** > > Just some notes... I do not have an OSPF interface defined for > Hipersockets. When I cancel the OMPROUTE started task, I am unable to > ping or reach other z/OS hipersocket addresses. I do have sourceVIPA > turned on... would this be the issue? At this point, I'm not too sure > why its not routing via the static route. > > IPCONFIG IGNOREREDIRECT > NOSYSPLEXROUTING > MULTIPATH > DATAGRAMFWD > SOURCEVIPA > PATHMTUDISC > IPSECURITY > ; > Beginroutes > route 5.5.5.0 255.255.255.0 = HSLINK28 MTU 57344 NODELAYA > Endroutes > ; > PRIMARYINTERFACE VLNKOS28 > > > Omproute options: > > AS_BOUNDARY_ROUTING > Import_RIP_Routes=No > Import_Static_Routes=No > Import_Direct_Routes=No > Import_Subnet_Routes=No > Originate_Default_Route=No > Originate_as_Type=2 > Default_Route_Cost=1 > > > - A traceroute to other z/OS LPARs yield something different: > > CS V1R13: Traceroute to 5.5.5.12 (5.5.5.12): > 1 5.5.5.12 (5.5.5.12) 3 ms 2 ms 0 ms > *** > > - From zLinux and z/VM, it looks right, where the first hop is by way > of the Hipersockets: > > [root@lsysg01a ~]# traceroute 5.5.5.12 traceroute to 5.5.5.12 > (5.5.5.12), 30 hops max, 60 byte packets > 1 5.5.5.12 (5.5.5.12) 0.645 ms 0.644 ms 0.724 ms > > Ready; T=0.01/0.01 07:54:24 > Trace route to 5.5.5.12 (5.5.5.12) > 1 (5.5.5.12) 10 ms 1 ms 0 ms > > > > > Serve DC is proud to present NeighborGood, a new, free tool to help > residents engage in meaningful service and connect with the causes and > organizations they care about. Visit NeighborGood at < > http://serve.dc.gov/service/neighborgood> > > > > > > -Original Message- > From: Linux on 390 Port
Re: Hipersockets between RHEL and z/OS
EZZ2350I MVS TCP/IP NETSTAT CS V1R13 TCPIP Name: TCPIP EZZ2700I Home address list: EZZ2701I Address Link Flg EZZ2703I 10.27.42.1 VLNKOS42 P EZZ2703I 10.82.10.13 LOSA42P0 EZZ2703I 10.82.10.81 LOSA43P0 EZZ2703I 5.5.5.42 HSLINK42 EZZ2703I 127.0.0.1LOOPBACK What would you like to see in the renovated Martin Luther King Jr. Memorial Library? Share your thoughts on the library's spaces and services at 'DC Public Librarys <http://www.dclibrary.org/> online idea community at dclibrary.ideascale.com <http://www.dclibrary.ideascale.com> -Original Message- From: Diep, David (OCTO-Contractor) Sent: Monday, November 04, 2013 12:01 PM To: LINUX-390@VM.MARIST.EDU Subject: RE: Hipersockets between RHEL and z/OS Offer, I have a test system... let me give it a shot. Right now, I know for a fact that my HS Home IP address is the last one listed. Thanks, David Diep -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Offer Baruch Sent: Monday, November 04, 2013 11:48 AM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets between RHEL and z/OS Hi You might be correct and this is a sourcevipa issue but i cant tell as i am lacking some data... Not many know this but the order of the home statment matters... you must place your hiper sockets interface before your vipa interface if you dont want that vipa to be applied to that interface. If you send the output of the following tso commands we can know for sure: Tso netstat home Tso netstat route That should give us the answer i think... Offer On Nov 4, 2013 3:11 PM, "Diep, David (OCTO-Contractor)" wrote: > Hi Offer, > > All IPtables are off, both for v4 and v6. > > Pinging from z/OS to zLinux does not work, nor does traceroute yield > anything. With all this said, I think it has something to do with the > way I have routing set up on the z/OS side. For all OSA traffic, I use > OSPF while for Hipersockets, I use static routing. > > - The reason I think this is because when I do a traceroute from z/OS > to z/VM, it looks like the route is by way of my VLAN through the > OSAs, though I have defined it for all traffic to 5.5.5.0/24 to be > routed through the Hipersockets link: > > CS V1R13: Traceroute to 5.5.5.2 (5.5.5.2): > 1 10.82.10.44 (10.82.10.44) 6 ms 10.82.10.121 (10.82.10.121) 0 ms > 10.82.10.44 > (10.82.10.44) 0 ms > *** > > Just some notes... I do not have an OSPF interface defined for > Hipersockets. When I cancel the OMPROUTE started task, I am unable to > ping or reach other z/OS hipersocket addresses. I do have sourceVIPA > turned on... would this be the issue? At this point, I'm not too sure > why its not routing via the static route. > > IPCONFIG IGNOREREDIRECT > NOSYSPLEXROUTING > MULTIPATH > DATAGRAMFWD > SOURCEVIPA > PATHMTUDISC > IPSECURITY > ; > Beginroutes > route 5.5.5.0 255.255.255.0 = HSLINK28 MTU 57344 NODELAYA > Endroutes > ; > PRIMARYINTERFACE VLNKOS28 > > > Omproute options: > > AS_BOUNDARY_ROUTING > Import_RIP_Routes=No > Import_Static_Routes=No > Import_Direct_Routes=No > Import_Subnet_Routes=No > Originate_Default_Route=No > Originate_as_Type=2 > Default_Route_Cost=1 > > > - A traceroute to other z/OS LPARs yield something different: > > CS V1R13: Traceroute to 5.5.5.12 (5.5.5.12): > 1 5.5.5.12 (5.5.5.12) 3 ms 2 ms 0 ms > *** > > - From zLinux and z/VM, it looks right, where the first hop is by way > of the Hipersockets: > > [root@lsysg01a ~]# traceroute 5.5.5.12 traceroute to 5.5.5.12 > (5.5.5.12), 30 hops max, 60 byte packets > 1 5.5.5.12 (5.5.5.12) 0.645 ms 0.644 ms 0.724 ms > > Ready; T=0.01/0.01 07:54:24 > Trace route to 5.5.5.12 (5.5.5.12) > 1 (5.5.5.12) 10 ms 1 ms 0 ms > > > > > Serve DC is proud to present NeighborGood, a new, free tool to help > residents engage in meaningful service and connect with the causes and > organizations they care about. Visit NeighborGood at < > http://serve.dc.gov/service/neighborgood> > > > > > > -Original Message- > From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of > Offer Baruch > Sent: Saturday, November 02, 2013 1:55 AM > To: LINUX-390@VM.MARIST.EDU > Subject: Re: Hipersockets between RHEL and z/OS > > Hi, > > First thing i would do is make sure your iptables service is stopped > and that the ipv6 iptables is down as well... > > Can you send the output of netstat route on z/os and route on linux? > > Also does pinging the z/os from linux works??? That what i got from > your first email... > > Offer Baruch >
Re: Hipersockets between RHEL and z/OS
Offer, I have a test system... let me give it a shot. Right now, I know for a fact that my HS Home IP address is the last one listed. Thanks, David Diep What would you like to see in the renovated Martin Luther King Jr. Memorial Library? Share your thoughts on the library's spaces and services at 'DC Public Librarys <http://www.dclibrary.org/> online idea community at dclibrary.ideascale.com <http://www.dclibrary.ideascale.com> -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Offer Baruch Sent: Monday, November 04, 2013 11:48 AM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets between RHEL and z/OS Hi You might be correct and this is a sourcevipa issue but i cant tell as i am lacking some data... Not many know this but the order of the home statment matters... you must place your hiper sockets interface before your vipa interface if you dont want that vipa to be applied to that interface. If you send the output of the following tso commands we can know for sure: Tso netstat home Tso netstat route That should give us the answer i think... Offer On Nov 4, 2013 3:11 PM, "Diep, David (OCTO-Contractor)" wrote: > Hi Offer, > > All IPtables are off, both for v4 and v6. > > Pinging from z/OS to zLinux does not work, nor does traceroute yield > anything. With all this said, I think it has something to do with the > way I have routing set up on the z/OS side. For all OSA traffic, I use > OSPF while for Hipersockets, I use static routing. > > - The reason I think this is because when I do a traceroute from z/OS > to z/VM, it looks like the route is by way of my VLAN through the > OSAs, though I have defined it for all traffic to 5.5.5.0/24 to be > routed through the Hipersockets link: > > CS V1R13: Traceroute to 5.5.5.2 (5.5.5.2): > 1 10.82.10.44 (10.82.10.44) 6 ms 10.82.10.121 (10.82.10.121) 0 ms > 10.82.10.44 > (10.82.10.44) 0 ms > *** > > Just some notes... I do not have an OSPF interface defined for > Hipersockets. When I cancel the OMPROUTE started task, I am unable to > ping or reach other z/OS hipersocket addresses. I do have sourceVIPA > turned on... would this be the issue? At this point, I'm not too sure > why its not routing via the static route. > > IPCONFIG IGNOREREDIRECT > NOSYSPLEXROUTING > MULTIPATH > DATAGRAMFWD > SOURCEVIPA > PATHMTUDISC > IPSECURITY > ; > Beginroutes > route 5.5.5.0 255.255.255.0 = HSLINK28 MTU 57344 NODELAYA > Endroutes > ; > PRIMARYINTERFACE VLNKOS28 > > > Omproute options: > > AS_BOUNDARY_ROUTING > Import_RIP_Routes=No > Import_Static_Routes=No > Import_Direct_Routes=No > Import_Subnet_Routes=No > Originate_Default_Route=No > Originate_as_Type=2 > Default_Route_Cost=1 > > > - A traceroute to other z/OS LPARs yield something different: > > CS V1R13: Traceroute to 5.5.5.12 (5.5.5.12): > 1 5.5.5.12 (5.5.5.12) 3 ms 2 ms 0 ms > *** > > - From zLinux and z/VM, it looks right, where the first hop is by way > of the Hipersockets: > > [root@lsysg01a ~]# traceroute 5.5.5.12 traceroute to 5.5.5.12 > (5.5.5.12), 30 hops max, 60 byte packets > 1 5.5.5.12 (5.5.5.12) 0.645 ms 0.644 ms 0.724 ms > > Ready; T=0.01/0.01 07:54:24 > Trace route to 5.5.5.12 (5.5.5.12) > 1 (5.5.5.12) 10 ms 1 ms 0 ms > > > > > Serve DC is proud to present NeighborGood, a new, free tool to help > residents engage in meaningful service and connect with the causes and > organizations they care about. Visit NeighborGood at < > http://serve.dc.gov/service/neighborgood> > > > > > > -Original Message- > From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of > Offer Baruch > Sent: Saturday, November 02, 2013 1:55 AM > To: LINUX-390@VM.MARIST.EDU > Subject: Re: Hipersockets between RHEL and z/OS > > Hi, > > First thing i would do is make sure your iptables service is stopped > and that the ipv6 iptables is down as well... > > Can you send the output of netstat route on z/os and route on linux? > > Also does pinging the z/os from linux works??? That what i got from > your first email... > > Offer Baruch > On Nov 1, 2013 7:06 PM, "Diep, David (OCTO-Contractor)" > > > wrote: > > > Hi Everyone, > > > > > > > > I got a good one (I think)... did I imagine this working at one point? > > > > > > > > I am not able to connect from my z/OS Hipersockets to my RHEL6.2 > > Hipersockets. > > > > > > > > I am able to successfully initiate a connection from my RHEL6.2 > > Hipersockets to my z/OS Hipe
Re: Hipersockets between RHEL and z/OS
Hi You might be correct and this is a sourcevipa issue but i cant tell as i am lacking some data... Not many know this but the order of the home statment matters... you must place your hiper sockets interface before your vipa interface if you dont want that vipa to be applied to that interface. If you send the output of the following tso commands we can know for sure: Tso netstat home Tso netstat route That should give us the answer i think... Offer On Nov 4, 2013 3:11 PM, "Diep, David (OCTO-Contractor)" wrote: > Hi Offer, > > All IPtables are off, both for v4 and v6. > > Pinging from z/OS to zLinux does not work, nor does traceroute yield > anything. With all this said, I think it has something to do with the way I > have routing set up on the z/OS side. For all OSA traffic, I use OSPF while > for Hipersockets, I use static routing. > > - The reason I think this is because when I do a traceroute from z/OS to > z/VM, it looks like the route is by way of my VLAN through the OSAs, though > I have defined it for all traffic to 5.5.5.0/24 to be routed through the > Hipersockets link: > > CS V1R13: Traceroute to 5.5.5.2 (5.5.5.2): > 1 10.82.10.44 (10.82.10.44) 6 ms 10.82.10.121 (10.82.10.121) 0 ms > 10.82.10.44 > (10.82.10.44) 0 ms > *** > > Just some notes... I do not have an OSPF interface defined for > Hipersockets. When I cancel the OMPROUTE started task, I am unable to ping > or reach other z/OS hipersocket addresses. I do have sourceVIPA turned > on... would this be the issue? At this point, I'm not too sure why its not > routing via the static route. > > IPCONFIG IGNOREREDIRECT > NOSYSPLEXROUTING > MULTIPATH > DATAGRAMFWD > SOURCEVIPA > PATHMTUDISC > IPSECURITY > ; > Beginroutes > route 5.5.5.0 255.255.255.0 = HSLINK28 MTU 57344 NODELAYA > Endroutes > ; > PRIMARYINTERFACE VLNKOS28 > > > Omproute options: > > AS_BOUNDARY_ROUTING > Import_RIP_Routes=No > Import_Static_Routes=No > Import_Direct_Routes=No > Import_Subnet_Routes=No > Originate_Default_Route=No > Originate_as_Type=2 > Default_Route_Cost=1 > > > - A traceroute to other z/OS LPARs yield something different: > > CS V1R13: Traceroute to 5.5.5.12 (5.5.5.12): > 1 5.5.5.12 (5.5.5.12) 3 ms 2 ms 0 ms > *** > > - From zLinux and z/VM, it looks right, where the first hop is by way of > the Hipersockets: > > [root@lsysg01a ~]# traceroute 5.5.5.12 > traceroute to 5.5.5.12 (5.5.5.12), 30 hops max, 60 byte packets > 1 5.5.5.12 (5.5.5.12) 0.645 ms 0.644 ms 0.724 ms > > Ready; T=0.01/0.01 07:54:24 > Trace route to 5.5.5.12 (5.5.5.12) > 1 (5.5.5.12) 10 ms 1 ms 0 ms > > > > > Serve DC is proud to present NeighborGood, a new, free tool to help > residents engage in meaningful service and connect with the causes and > organizations they care about. Visit NeighborGood at < > http://serve.dc.gov/service/neighborgood> > > > > > > -Original Message- > From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of > Offer Baruch > Sent: Saturday, November 02, 2013 1:55 AM > To: LINUX-390@VM.MARIST.EDU > Subject: Re: Hipersockets between RHEL and z/OS > > Hi, > > First thing i would do is make sure your iptables service is stopped and > that the ipv6 iptables is down as well... > > Can you send the output of netstat route on z/os and route on linux? > > Also does pinging the z/os from linux works??? That what i got from your > first email... > > Offer Baruch > On Nov 1, 2013 7:06 PM, "Diep, David (OCTO-Contractor)" > > wrote: > > > Hi Everyone, > > > > > > > > I got a good one (I think)... did I imagine this working at one point? > > > > > > > > I am not able to connect from my z/OS Hipersockets to my RHEL6.2 > > Hipersockets. > > > > > > > > I am able to successfully initiate a connection from my RHEL6.2 > > Hipersockets to my z/OS Hipersockets. > > > > I am able to successfully initiate a connection from my RHEL6.2 > > Hipersockets to my z/VM Hipersockets. > > > > I am able to successfully initiate a connection from my z/OS > > Hipersockets to my z/VM Hipersockets. > > > > I am able to successfully initiate a connection from my z/VM > > Hipersockets to my z/OS Hipersockets. > > > > I am able to successfully initiate a connection from my z/VM > > Hipersockets to my RHEL6.2 Hipersockets. > > > > > > > > What gives? > > > > > > > > Hipersockets on zLinux: > > > > [root@vipservd ~]# cat /etc/sys
Re: Hipersockets between RHEL and z/OS
Hi Offer, All IPtables are off, both for v4 and v6. Pinging from z/OS to zLinux does not work, nor does traceroute yield anything. With all this said, I think it has something to do with the way I have routing set up on the z/OS side. For all OSA traffic, I use OSPF while for Hipersockets, I use static routing. - The reason I think this is because when I do a traceroute from z/OS to z/VM, it looks like the route is by way of my VLAN through the OSAs, though I have defined it for all traffic to 5.5.5.0/24 to be routed through the Hipersockets link: CS V1R13: Traceroute to 5.5.5.2 (5.5.5.2): 1 10.82.10.44 (10.82.10.44) 6 ms 10.82.10.121 (10.82.10.121) 0 ms 10.82.10.44 (10.82.10.44) 0 ms *** Just some notes... I do not have an OSPF interface defined for Hipersockets. When I cancel the OMPROUTE started task, I am unable to ping or reach other z/OS hipersocket addresses. I do have sourceVIPA turned on... would this be the issue? At this point, I'm not too sure why its not routing via the static route. IPCONFIG IGNOREREDIRECT NOSYSPLEXROUTING MULTIPATH DATAGRAMFWD SOURCEVIPA PATHMTUDISC IPSECURITY ; Beginroutes route 5.5.5.0 255.255.255.0 = HSLINK28 MTU 57344 NODELAYA Endroutes ; PRIMARYINTERFACE VLNKOS28 Omproute options: AS_BOUNDARY_ROUTING Import_RIP_Routes=No Import_Static_Routes=No Import_Direct_Routes=No Import_Subnet_Routes=No Originate_Default_Route=No Originate_as_Type=2 Default_Route_Cost=1 - A traceroute to other z/OS LPARs yield something different: CS V1R13: Traceroute to 5.5.5.12 (5.5.5.12): 1 5.5.5.12 (5.5.5.12) 3 ms 2 ms 0 ms *** - From zLinux and z/VM, it looks right, where the first hop is by way of the Hipersockets: [root@lsysg01a ~]# traceroute 5.5.5.12 traceroute to 5.5.5.12 (5.5.5.12), 30 hops max, 60 byte packets 1 5.5.5.12 (5.5.5.12) 0.645 ms 0.644 ms 0.724 ms Ready; T=0.01/0.01 07:54:24 Trace route to 5.5.5.12 (5.5.5.12) 1 (5.5.5.12) 10 ms 1 ms 0 ms Serve DC is proud to present NeighborGood, a new, free tool to help residents engage in meaningful service and connect with the causes and organizations they care about. Visit NeighborGood at <http://serve.dc.gov/service/neighborgood> -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Offer Baruch Sent: Saturday, November 02, 2013 1:55 AM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets between RHEL and z/OS Hi, First thing i would do is make sure your iptables service is stopped and that the ipv6 iptables is down as well... Can you send the output of netstat route on z/os and route on linux? Also does pinging the z/os from linux works??? That what i got from your first email... Offer Baruch On Nov 1, 2013 7:06 PM, "Diep, David (OCTO-Contractor)" wrote: > Hi Everyone, > > > > I got a good one (I think)... did I imagine this working at one point? > > > > I am not able to connect from my z/OS Hipersockets to my RHEL6.2 > Hipersockets. > > > > I am able to successfully initiate a connection from my RHEL6.2 > Hipersockets to my z/OS Hipersockets. > > I am able to successfully initiate a connection from my RHEL6.2 > Hipersockets to my z/VM Hipersockets. > > I am able to successfully initiate a connection from my z/OS > Hipersockets to my z/VM Hipersockets. > > I am able to successfully initiate a connection from my z/VM > Hipersockets to my z/OS Hipersockets. > > I am able to successfully initiate a connection from my z/VM > Hipersockets to my RHEL6.2 Hipersockets. > > > > What gives? > > > > Hipersockets on zLinux: > > [root@vipservd ~]# cat /etc/sysconfig/network-scripts/ifcfg-hsi0 > > DEVICE="hsi0" > > BOOTPROTO="static" > > DNS1="10.201.1.241" > > IPADDR="5.5.5.57" > > NETMASK="255.255.255.0" > > NETTYPE="qeth" > > ONBOOT="yes" > > SUBCHANNELS="0.0.7400,0.0.7401,0.0.7402" > > TYPE="Ethernet" > > ARP="no" > > > > Hipersockets on z/OS: > > ; > > DEVICE IUTIQDED MPCIPA AUTORESTART > > LINK HSLINK16 IPAQIDIO IUTIQDED > > START IUTIQDED > > ; > > HOME > >5.5.5.16 HSLINK16 > > ; > > ; Begin / End route table > > ; > > Beginroutes > > route 5.5.5.0 255.255.255.0 = HSLINK16 mtu 57344 NODELAYA > > Endroutes > > > > > > Thanks!! > > > > David Diep > > Serve DC is proud to present NeighborGood, a new, free tool to help > residents engage in meaningful service and connect with the causes and > organizations they care about. Visit NeighborGood at > http://serve.dc.gov/servi
Re: Hipersockets between RHEL and z/OS
Hi, First thing i would do is make sure your iptables service is stopped and that the ipv6 iptables is down as well... Can you send the output of netstat route on z/os and route on linux? Also does pinging the z/os from linux works??? That what i got from your first email... Offer Baruch On Nov 1, 2013 7:06 PM, "Diep, David (OCTO-Contractor)" wrote: > Hi Everyone, > > > > I got a good one (I think)... did I imagine this working at one point? > > > > I am not able to connect from my z/OS Hipersockets to my RHEL6.2 > Hipersockets. > > > > I am able to successfully initiate a connection from my RHEL6.2 > Hipersockets to my z/OS Hipersockets. > > I am able to successfully initiate a connection from my RHEL6.2 > Hipersockets to my z/VM Hipersockets. > > I am able to successfully initiate a connection from my z/OS Hipersockets > to my z/VM Hipersockets. > > I am able to successfully initiate a connection from my z/VM Hipersockets > to my z/OS Hipersockets. > > I am able to successfully initiate a connection from my z/VM Hipersockets > to my RHEL6.2 Hipersockets. > > > > What gives? > > > > Hipersockets on zLinux: > > [root@vipservd ~]# cat /etc/sysconfig/network-scripts/ifcfg-hsi0 > > DEVICE="hsi0" > > BOOTPROTO="static" > > DNS1="10.201.1.241" > > IPADDR="5.5.5.57" > > NETMASK="255.255.255.0" > > NETTYPE="qeth" > > ONBOOT="yes" > > SUBCHANNELS="0.0.7400,0.0.7401,0.0.7402" > > TYPE="Ethernet" > > ARP="no" > > > > Hipersockets on z/OS: > > ; > > DEVICE IUTIQDED MPCIPA AUTORESTART > > LINK HSLINK16 IPAQIDIO IUTIQDED > > START IUTIQDED > > ; > > HOME > >5.5.5.16 HSLINK16 > > ; > > ; Begin / End route table > > ; > > Beginroutes > > route 5.5.5.0 255.255.255.0 = HSLINK16 mtu 57344 NODELAYA > > Endroutes > > > > > > Thanks!! > > > > David Diep > > Serve DC is proud to present NeighborGood, a new, free tool to help > residents engage in meaningful service and connect with the causes and > organizations they care about. Visit NeighborGood at > http://serve.dc.gov/service/neighborgood > > -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets between RHEL and z/OS
I use a host entry and then request that the apps teams make the calls to the FQDN in the host file. Many ways to solve this - this works for us. -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Diep, David (OCTO-Contractor) Sent: Friday, November 01, 2013 2:19 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets between RHEL and z/OS Emmit, That could very well be the case... I did not define a specific route for Hipersockets in RHEL (always wondered about that). As for my z/OS, I use OSPF for routing, but static routes for Hipersockets traffic. ; DEVICE IUTIQDED MPCIPA AUTORESTART LINK HSLINK16 IPAQIDIO IUTIQDED START IUTIQDED ; HOME 5.5.5.16 HSLINK16 ; ; Begin / End route table ; Beginroutes route 5.5.5.0 255.255.255.0 = HSLINK16 mtu 57344 NODELAYA Endroutes Have you guys had this issue? -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Emmett O'Grady Sent: Friday, November 01, 2013 2:05 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets between RHEL and z/OS Is the request going out your default route as opposed to the hipersocket? -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Diep, David (OCTO-Contractor) Sent: Friday, November 01, 2013 2:02 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets between RHEL and z/OS Hi J.B., It tells me this: EZA1450I IBM FTP CS V1R13 EZA1554I Connecting to: 5.5.5.57 port: 21. EZA2589E Connection to server interrupted or timed out. Initial connection *** CS V1R13: Pinging host 5.5.5.57 Ping #1 timed out *** Nothing in SYSLOGD. Serve DC is proud to present NeighborGood, a new, free tool to help residents engage in meaningful service and connect with the causes and organizations they care about. Visit NeighborGood at <http://serve.dc.gov/service/neighborgood> -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of J B Mills III Sent: Friday, November 01, 2013 1:43 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets between RHEL and z/OS What is the message that is telling you that you can not connect to your z/OS hipersocket? Regards, J.B. Mills III Solutions Architect 1515 Poydras Street Suite 1925 New Orleans, LA. 70112 Cell 614-226-0284 jmil...@us.ibm.com "The only difference between a good day and a bad day, is your attitude." -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ If you are not the addressee, please inform us immediately that you have received this e-mail by mistake, and delete it. We thank you for your support. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ If you are not the addressee, please inform us immediately that you have received this e-mail by mistake, and delete it. We thank you for your support. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets between RHEL and z/OS
Emmit, That could very well be the case... I did not define a specific route for Hipersockets in RHEL (always wondered about that). As for my z/OS, I use OSPF for routing, but static routes for Hipersockets traffic. ; DEVICE IUTIQDED MPCIPA AUTORESTART LINK HSLINK16 IPAQIDIO IUTIQDED START IUTIQDED ; HOME 5.5.5.16 HSLINK16 ; ; Begin / End route table ; Beginroutes route 5.5.5.0 255.255.255.0 = HSLINK16 mtu 57344 NODELAYA Endroutes Have you guys had this issue? -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Emmett O'Grady Sent: Friday, November 01, 2013 2:05 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets between RHEL and z/OS Is the request going out your default route as opposed to the hipersocket? -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Diep, David (OCTO-Contractor) Sent: Friday, November 01, 2013 2:02 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets between RHEL and z/OS Hi J.B., It tells me this: EZA1450I IBM FTP CS V1R13 EZA1554I Connecting to: 5.5.5.57 port: 21. EZA2589E Connection to server interrupted or timed out. Initial connection *** CS V1R13: Pinging host 5.5.5.57 Ping #1 timed out *** Nothing in SYSLOGD. Serve DC is proud to present NeighborGood, a new, free tool to help residents engage in meaningful service and connect with the causes and organizations they care about. Visit NeighborGood at <http://serve.dc.gov/service/neighborgood> -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of J B Mills III Sent: Friday, November 01, 2013 1:43 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets between RHEL and z/OS What is the message that is telling you that you can not connect to your z/OS hipersocket? Regards, J.B. Mills III Solutions Architect 1515 Poydras Street Suite 1925 New Orleans, LA. 70112 Cell 614-226-0284 jmil...@us.ibm.com "The only difference between a good day and a bad day, is your attitude." -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ If you are not the addressee, please inform us immediately that you have received this e-mail by mistake, and delete it. We thank you for your support. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets between RHEL and z/OS
Is the request going out your default route as opposed to the hipersocket? -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Diep, David (OCTO-Contractor) Sent: Friday, November 01, 2013 2:02 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets between RHEL and z/OS Hi J.B., It tells me this: EZA1450I IBM FTP CS V1R13 EZA1554I Connecting to: 5.5.5.57 port: 21. EZA2589E Connection to server interrupted or timed out. Initial connection *** CS V1R13: Pinging host 5.5.5.57 Ping #1 timed out *** Nothing in SYSLOGD. Serve DC is proud to present NeighborGood, a new, free tool to help residents engage in meaningful service and connect with the causes and organizations they care about. Visit NeighborGood at <http://serve.dc.gov/service/neighborgood> -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of J B Mills III Sent: Friday, November 01, 2013 1:43 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets between RHEL and z/OS What is the message that is telling you that you can not connect to your z/OS hipersocket? Regards, J.B. Mills III Solutions Architect 1515 Poydras Street Suite 1925 New Orleans, LA. 70112 Cell 614-226-0284 jmil...@us.ibm.com "The only difference between a good day and a bad day, is your attitude." -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ If you are not the addressee, please inform us immediately that you have received this e-mail by mistake, and delete it. We thank you for your support. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets between RHEL and z/OS
Hi J.B., It tells me this: EZA1450I IBM FTP CS V1R13 EZA1554I Connecting to: 5.5.5.57 port: 21. EZA2589E Connection to server interrupted or timed out. Initial connection *** CS V1R13: Pinging host 5.5.5.57 Ping #1 timed out *** Nothing in SYSLOGD. Serve DC is proud to present NeighborGood, a new, free tool to help residents engage in meaningful service and connect with the causes and organizations they care about. Visit NeighborGood at <http://serve.dc.gov/service/neighborgood> -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of J B Mills III Sent: Friday, November 01, 2013 1:43 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets between RHEL and z/OS What is the message that is telling you that you can not connect to your z/OS hipersocket? Regards, J.B. Mills III Solutions Architect 1515 Poydras Street Suite 1925 New Orleans, LA. 70112 Cell 614-226-0284 jmil...@us.ibm.com "The only difference between a good day and a bad day, is your attitude." -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets between RHEL and z/OS
What is the message that is telling you that you can not connect to your z/OS hipersocket? Regards, J.B. Mills III Solutions Architect 1515 Poydras Street Suite 1925 New Orleans, LA. 70112 Cell 614-226-0284 jmil...@us.ibm.com "The only difference between a good day and a bad day, is your attitude." -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets "broke" Ethernet?
We had this same problem 2 years ago. This is what we did to solve it: Be sure to place the hipersocket address first in the home list before any SOURCEVIPA addresses in PROFILE TCPIP fm. Do this in z/VM and in z/OS both. Kyle Stewart -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Chase, John Sent: Wednesday, February 27, 2013 1:34 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets "broke" Ethernet? > -Original Message- > From: Chase, John > > > [ snip ] > > Now, what's missing from this file that is present in the same file on > our "golden" image is the > statement: > > OPTIONS="layer2=0 portno=0" > > I don't know why that statement is missing, unless our "network guy" > deleted it from the penguin that doesn't talk while he was trying to figure > out Hipersockets. > > I also notice the "golden" copy has BOOTPROTO="static" instead of "none". > > I'll set both back to "golden" and see what happens. And that fixed that problem. We now have both interfaces working, but the hipersockets network appears to be "one-way" at the moment: We can ping z/OS from Linux, but cannot ping Linux from z/OS. I don't believe that side of the question belongs on this forum, but if somebody wants to respond off-list, please do. -jc- ** Information contained in this e-mail message and in any attachments thereto is confidential. If you are not the intended recipient, please destroy this message, delete any copies held on your systems, notify the sender immediately, and refrain from using or disclosing all or any part of its content to any other person. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ == THIS ELECTRONIC MESSAGE, INCLUDING ANY ACCOMPANYING DOCUMENTS, IS CONFIDENTIAL and may contain information that is privileged and exempt from disclosure under applicable law. If you are neither the intended recipient nor responsible for delivering the message to the intended recipient, please note that any dissemination, distribution, copying or the taking of any action in reliance upon the message is strictly prohibited. If you have received this communication in error, please notify the sender immediately. Thank you. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets "broke" Ethernet?
Be sure to place the hipersocket address first in the home list before any SOURCEVIPA addresses in PROFILE TCPIP fm. Do this in z/VM and in z/OS both. Kyle Stewart -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Chase, John Sent: Wednesday, February 27, 2013 1:34 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets "broke" Ethernet? > -Original Message- > From: Chase, John > > > [ snip ] > > Now, what's missing from this file that is present in the same file on > our "golden" image is the > statement: > > OPTIONS="layer2=0 portno=0" > > I don't know why that statement is missing, unless our "network guy" > deleted it from the penguin that doesn't talk while he was trying to figure > out Hipersockets. > > I also notice the "golden" copy has BOOTPROTO="static" instead of "none". > > I'll set both back to "golden" and see what happens. And that fixed that problem. We now have both interfaces working, but the hipersockets network appears to be "one-way" at the moment: We can ping z/OS from Linux, but cannot ping Linux from z/OS. I don't believe that side of the question belongs on this forum, but if somebody wants to respond off-list, please do. -jc- ** Information contained in this e-mail message and in any attachments thereto is confidential. If you are not the intended recipient, please destroy this message, delete any copies held on your systems, notify the sender immediately, and refrain from using or disclosing all or any part of its content to any other person. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ == THIS ELECTRONIC MESSAGE, INCLUDING ANY ACCOMPANYING DOCUMENTS, IS CONFIDENTIAL and may contain information that is privileged and exempt from disclosure under applicable law. If you are neither the intended recipient nor responsible for delivering the message to the intended recipient, please note that any dissemination, distribution, copying or the taking of any action in reliance upon the message is strictly prohibited. If you have received this communication in error, please notify the sender immediately. Thank you. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets "broke" Ethernet?
> -Original Message- > From: Chase, John > > > [ snip ] > > Now, what's missing from this file that is present in the same file on our > "golden" image is the > statement: > > OPTIONS="layer2=0 portno=0" > > I don't know why that statement is missing, unless our "network guy" deleted > it from the penguin that > doesn't talk while he was trying to figure out Hipersockets. > > I also notice the "golden" copy has BOOTPROTO="static" instead of "none". > > I'll set both back to "golden" and see what happens. And that fixed that problem. We now have both interfaces working, but the hipersockets network appears to be "one-way" at the moment: We can ping z/OS from Linux, but cannot ping Linux from z/OS. I don't believe that side of the question belongs on this forum, but if somebody wants to respond off-list, please do. -jc- ** Information contained in this e-mail message and in any attachments thereto is confidential. If you are not the intended recipient, please destroy this message, delete any copies held on your systems, notify the sender immediately, and refrain from using or disclosing all or any part of its content to any other person. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets "broke" Ethernet?
> -Original Message- > From: Linux on 390 Port On Behalf Of Lee Stewart > > RHEL has an error where if you use system-config-network on a layer 2 > interface it "forgets" that it > was layer 2. Reported a year ago as Bug > 809534 - system-config-network ignores option field. No fix. > > Did you use system-config-network? We did not see a directory or file with that exact name, so I'll guess we did not. One file named network, in /etc/sysconfig, contains only three lines: NETWORKING=yes HOSTAME=myPenguin.ussco.com GATEWAY=10.1.1.1 The ifcfg-eth0 cited earlier lives in /etc/sysconfig/network-scripts. In case I forgot to mention it, we're working with RHEL 6.3. -jc- [ snip ] ** Information contained in this e-mail message and in any attachments thereto is confidential. If you are not the intended recipient, please destroy this message, delete any copies held on your systems, notify the sender immediately, and refrain from using or disclosing all or any part of its content to any other person. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets "broke" Ethernet?
RHEL has an error where if you use system-config-network on a layer 2 interface it "forgets" that it was layer 2. Reported a year ago as Bug 809534 - system-config-network ignores option field. No fix. Did you use system-config-network? Lee On 2/26/2013 6:30 PM, Chase, John wrote: -Original Message- From: Linux on 390 Port On Behalf Of Mark Post On 2/26/2013 at 06:08 PM, "Chase, John" wrote: -snip- qeth: register layer 3 discipline This line, and the line 2 down is interesting. It looks like the NIC is trying to be both Layer 3 and Layer 2. Which is it _supposed_ to be? I believe it's supposed to be Layer 2; connects to a vswitch. qdio: 0.0.7002 HS on SC 2 using AI:1 QEBSM:1 PCI:0 TDD:1 SIGA:RW AO qeth: register layer 2 discipline What all does the ifcfg-eth0 file have in it? [root@myPenguin ~]# cat ifcfg-eth0 DEVICE=eth0 BOOTPROTO=none DNS1=10.1.4.73 DOMAIN=ourshop.com GATEWAY=10.1.1.1 IPADDR=10.1.1.36 MTU=1500 NETMASK=255.255.255.192 NETTYPE=qeth NM_CONTROLLED=yes ONBOOT=yes PORTNAME="DONTCARE" SUBCHANNELS="0.0.0600,0.0.0601,0.0.0602" TYPE=Ethernet UUID="646a685e-75d1-4ef4-bc5e-7482b66a7981" IPV6INIT=no USERCTL=no [root@myPenguin ~]# Now, what's missing from this file that is present in the same file on our "golden" image is the statement: OPTIONS="layer2=0 portno=0" I don't know why that statement is missing, unless our "network guy" deleted it from the penguin that doesn't talk while he was trying to figure out Hipersockets. I also notice the "golden" copy has BOOTPROTO="static" instead of "none". I'll set both back to "golden" and see what happens. -jc- ** Information contained in this e-mail message and in any attachments thereto is confidential. If you are not the intended recipient, please destroy this message, delete any copies held on your systems, notify the sender immediately, and refrain from using or disclosing all or any part of its content to any other person. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- Lee Stewart, Senior SE Sirius Computer Solutions Phone: (303) 996-7122 Email: lee.stew...@siriuscom.com Web: www.siriuscom.com -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets "broke" Ethernet?
> -Original Message- > From: Linux on 390 Port On Behalf Of Mark Post > > >>> On 2/26/2013 at 06:08 PM, "Chase, John" wrote: > -snip- > > > qeth: register layer 3 discipline > > This line, and the line 2 down is interesting. It looks like the NIC is > trying to be both Layer 3 and > Layer 2. Which is it _supposed_ to be? I believe it's supposed to be Layer 2; connects to a vswitch. > > qdio: 0.0.7002 HS on SC 2 using AI:1 QEBSM:1 PCI:0 TDD:1 SIGA:RW AO > > qeth: register layer 2 discipline > > What all does the ifcfg-eth0 file have in it? [root@myPenguin ~]# cat ifcfg-eth0 DEVICE=eth0 BOOTPROTO=none DNS1=10.1.4.73 DOMAIN=ourshop.com GATEWAY=10.1.1.1 IPADDR=10.1.1.36 MTU=1500 NETMASK=255.255.255.192 NETTYPE=qeth NM_CONTROLLED=yes ONBOOT=yes PORTNAME="DONTCARE" SUBCHANNELS="0.0.0600,0.0.0601,0.0.0602" TYPE=Ethernet UUID="646a685e-75d1-4ef4-bc5e-7482b66a7981" IPV6INIT=no USERCTL=no [root@myPenguin ~]# Now, what's missing from this file that is present in the same file on our "golden" image is the statement: OPTIONS="layer2=0 portno=0" I don't know why that statement is missing, unless our "network guy" deleted it from the penguin that doesn't talk while he was trying to figure out Hipersockets. I also notice the "golden" copy has BOOTPROTO="static" instead of "none". I'll set both back to "golden" and see what happens. -jc- ** Information contained in this e-mail message and in any attachments thereto is confidential. If you are not the intended recipient, please destroy this message, delete any copies held on your systems, notify the sender immediately, and refrain from using or disclosing all or any part of its content to any other person. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets "broke" Ethernet?
How are 7000-7002 defined? Seeing the directory entry would be useful too. An error is happening before the ipl moment. David Kreuter Original message From: Mark Post Date: To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets "broke" Ethernet? >>> On 2/26/2013 at 06:08 PM, "Chase, John" wrote: -snip- > Tried an "interactive" IPL and startup of the Linux vm, and saw some error > messages that indicate we've got some misteaks [sic] in the definitions of > the hipersockets virtual devices (but after the guest is started, ifconfig > shows only the hipersockets interface as active). > > Here's the VM logon of the guest: > > LOGON AT 15:40:29 CST TUESDAY 02/26/13  > Command complete > NIC 0600 is created; devices 0600-0602 defined   > NIC 0600 is connected to VSWITCH SYSTEM VMBSW1  > HCPNIC2781E NIC 7000 not created; network devices 7000-7002 could not be > defined > z/VM V6.2.0   2013-02-04 13:33 > > Vdevs 7000-7002 are for the hipersockets interface. The NIC at vaddr 600 is > the Ethernet interface that previously worked just fine. > > Here's the part of the Linux console log pertaining to qeth: > > qeth: loading core functions > vmur: z/VM virtual unit record device driver loaded. > NET: Registered protocol family 10  > lo: Disabled Privacy Extensions > qeth: register layer 3 discipline This line, and the line 2 down is interesting. It looks like the NIC is trying to be both Layer 3 and Layer 2. Which is it _supposed_ to be? > qdio: 0.0.7002 HS on SC 2 using AI:1 QEBSM:1 PCI:0 TDD:1 SIGA:RW AO > qeth: register layer 2 discipline What all does the ifcfg-eth0 file have in it? > qeth 0.0.0600: The qeth device is not configured for the OSI layer required > by z > /VM > qeth 0.0.0600: The qeth device driver failed to recover an error on the > device > qeth: irb : 00 c2 60 17 0d b9 30 38 0e 00 10 00 00 80 00 00 > ..`...08... > .   > qeth: irb 0010: 01 02 00 00 00 00 00 00 00 00 00 00 00 00 00 00 > ... > .   > qeth: sense data : 02 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 > > > qeth: sense data 0010: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 > > > qeth 0.0.0600: The qeth device driver failed to recover an error on the > device Mark Post -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets "broke" Ethernet?
>>> On 2/26/2013 at 06:08 PM, "Chase, John" wrote: -snip- > Tried an "interactive" IPL and startup of the Linux vm, and saw some error > messages that indicate we've got some misteaks [sic] in the definitions of > the hipersockets virtual devices (but after the guest is started, ifconfig > shows only the hipersockets interface as active). > > Here's the VM logon of the guest: > > LOGON AT 15:40:29 CST TUESDAY 02/26/13 > Command complete > NIC 0600 is created; devices 0600-0602 defined > NIC 0600 is connected to VSWITCH SYSTEM VMBSW1 > HCPNIC2781E NIC 7000 not created; network devices 7000-7002 could not be > defined > z/VM V6.2.02013-02-04 13:33 > > Vdevs 7000-7002 are for the hipersockets interface. The NIC at vaddr 600 is > the Ethernet interface that previously worked just fine. > > Here's the part of the Linux console log pertaining to qeth: > > qeth: loading core functions > vmur: z/VM virtual unit record device driver loaded. > NET: Registered protocol family 10 > lo: Disabled Privacy Extensions > qeth: register layer 3 discipline This line, and the line 2 down is interesting. It looks like the NIC is trying to be both Layer 3 and Layer 2. Which is it _supposed_ to be? > qdio: 0.0.7002 HS on SC 2 using AI:1 QEBSM:1 PCI:0 TDD:1 SIGA:RW AO > qeth: register layer 2 discipline What all does the ifcfg-eth0 file have in it? > qeth 0.0.0600: The qeth device is not configured for the OSI layer required > by z > /VM > qeth 0.0.0600: The qeth device driver failed to recover an error on the > device > qeth: irb : 00 c2 60 17 0d b9 30 38 0e 00 10 00 00 80 00 00 > ..`...08... > . > qeth: irb 0010: 01 02 00 00 00 00 00 00 00 00 00 00 00 00 00 00 > ... > . > qeth: sense data : 02 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 > > > qeth: sense data 0010: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 > > > qeth 0.0.0600: The qeth device driver failed to recover an error on the > device Mark Post -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets "broke" Ethernet?
> -Original Message- > From: Linux on 390 Port On Behalf Of Mark Post > > >>> On 2/26/2013 at 04:48 PM, "Chase, John" wrote: > > How do we configure Linux so that both interfaces (eth0 and hsi0) are > > configured at startup time? > > Look at /etc/sysconfig/network-scripts/ifcfg-eth0 to see what is in there for > ONBOOT. For the > interface to be started automatically, it should be set to "yes." Thanks. Checked; it's set to "yes". Tried an "interactive" IPL and startup of the Linux vm, and saw some error messages that indicate we've got some misteaks [sic] in the definitions of the hipersockets virtual devices (but after the guest is started, ifconfig shows only the hipersockets interface as active). Here's the VM logon of the guest: LOGON AT 15:40:29 CST TUESDAY 02/26/13 Command complete NIC 0600 is created; devices 0600-0602 defined NIC 0600 is connected to VSWITCH SYSTEM VMBSW1 HCPNIC2781E NIC 7000 not created; network devices 7000-7002 could not be defined z/VM V6.2.02013-02-04 13:33 Vdevs 7000-7002 are for the hipersockets interface. The NIC at vaddr 600 is the Ethernet interface that previously worked just fine. Here's the part of the Linux console log pertaining to qeth: qeth: loading core functions vmur: z/VM virtual unit record device driver loaded. NET: Registered protocol family 10 lo: Disabled Privacy Extensions qeth: register layer 3 discipline qdio: 0.0.7002 HS on SC 2 using AI:1 QEBSM:1 PCI:0 TDD:1 SIGA:RW AO qeth: register layer 2 discipline qeth 0.0.0600: The qeth device is not configured for the OSI layer required by z /VM qeth 0.0.0600: The qeth device driver failed to recover an error on the device qeth: irb : 00 c2 60 17 0d b9 30 38 0e 00 10 00 00 80 00 00 ..`...08... . qeth: irb 0010: 01 02 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ... . qeth: sense data : 02 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 qeth: sense data 0010: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 qeth 0.0.0600: The qeth device driver failed to recover an error on the device qeth 0.0.7000: Device is a HiperSockets card (level: HSGR) with link type HiperSockets. qeth 0.0.7000: Hardware IP fragmentation not supported on hsi0 qeth 0.0.7000: Inbound source MAC-address not supported on hsi0 qeth 0.0.7000: VLAN enabled qeth 0.0.7000: Multicast enabled qeth 0.0.7000: IPV6 enabled qeth 0.0.7000: Broadcast enabled qeth 0.0.7000: Using SW checksumming on hsi0. qeth 0.0.7000: Outbound TSO not supported on hsi0 [ OK ] Setting hostname : [ OK ] Bringing up interface eth0: WARNING: /etc/modprobe.conf line 1: ignoring bad li ne starting with 'hsi0' WARNING: Deprecated config file /etc/modprobe.conf, all config files belong into /etc/modprobe.d/. WARNING: /etc/modprobe.conf line 1: ignoring bad line starting with 'hsi0' WARNING: Deprecated config file /etc/modprobe.conf, all config files belong into /etc/modprobe.d/. Device eth0 does not seem to be present, delaying initialization. [FAILED] Bringing up interface hsi0: [ OK ] The modprobe.conf file was created by a colleague, and contains one line: 'hsi0 alias qeth' (without the apostrophes). That's all the diagnostics I have at the moment. -jc- ** Information contained in this e-mail message and in any attachments thereto is confidential. If you are not the intended recipient, please destroy this message, delete any copies held on your systems, notify the sender immediately, and refrain from using or disclosing a
Re: Hipersockets "broke" Ethernet?
>>> On 2/26/2013 at 04:48 PM, "Chase, John" wrote: > How do we configure Linux so that both interfaces (eth0 and hsi0) are > configured at startup time? Look at /etc/sysconfig/network-scripts/ifcfg-eth0 to see what is in there for ONBOOT. For the interface to be started automatically, it should be set to "yes." Mark Post -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets over Guest LAN to MVS
On Tue, Jul 17, 2012 at 7:50 PM, Alan Altmark wrote: > > Perhaps there is a misunderstanding. I have no objection to real > HiperSockets for backups (or any other reason) as long as you realize that > HiperSockets consume z CPU cycles since it is the CPU that moves the data, > not the IOPs. > > Anything that moves packets will cost CPU cycles, in the Linux TCP/IP stack and in CP. The biggest factor is the number of packets, so 56K packets on HiperSockets is hard to beat in comparison with 1500 byte on the LAN. A virtual NIC (connected to a VSWITCH or Guest LAN) is some 50% more expensive than using HiperSockets. When using a Linux guest as virtual router, things do add up rather fast (packets are handled 3 times, plus the cost of the virtual router). Whether these aspects should play a major role in your network design really depends on the data volume. Many servers don't transfer hundreds of GB's per day. I would suggest to use a network design that makes sense and measure your CPU cost and network volume to see whether you need to connect some of the busy guests differently. Rob -- Rob van der Heij Velocity Software http://www.velocitysoftware.com/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets over Guest LAN to MVS
On Tuesday, 07/17/2012 at 11:34 EDT, "Norris, Chet" wrote: > It sounds like Alan's "HiperSocket VSWITCH Bridge" could be an option, but > multiple people (including Alan) have recommended using direct Hipersocket > device definitions for each zLinux user. Perhaps there is a misunderstanding. I have no objection to real HiperSockets for backups (or any other reason) as long as you realize that HiperSockets consume z CPU cycles since it is the CPU that moves the data, not the IOPs. > If I make it simple and define unique devices to each user, how do I fail over > to a different device if MVS moves to another machine? I'm also going to be > limited as to the number of zLinux users on any given CHPID, or is that number > high enough that its not an issue? I should have been paying closer attention. The HiperSocket VSWITCH bridge is specifically designed to give you the performance benefit of real HiperSockets when the target is in a neighboring LPAR, but the flexibility to transparently use ethernets when the target is elsewhere. There's no longer any need to configure a secondary paths in order to accommodate the ripping noise you hear when HiperSockets partners are torn out of the CEC (e.g. DR or failover configs). BUT, the bridge only works with MVS when the bridge is configured to use the IEDN in a zEnterprise ensemble (OSX chpids). For non-ensemble configurations (OSD chpids), z/OS drives the HiperSocket in a way that prevents use of the bridge. Also, be aware that the bridge is available only on a z196 or z114 with z/VM 6.2. Across all HiperSocket chpids, you have have 12K unit addresses. Since each host/guest uses 3 UAs, you can have up to 4096 hosts using HiperSockets on the CEC. Alan Altmark Senior Managing z/VM and Linux Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets over Guest LAN to MVS
I guess my main drive to use a common guest LAN was to avoid having to assign a unique device set for each zLinux guest by specifying a SPECIAL definition in each. The main use for the Hipersocket connections would be to run zLinux backup/restore processing based on the MVS platform, and I theorized that the MVS Hipersocket device would be a bottleneck anyway so why not use TCPIP as a centralized source. It sounds like using Guest LANs is the wrong answer, even though it would work. I really need to have this process available no matter which physical CEC MVS happens to be on, so I would need an alternative route for the Hipersocket devices. It sounds like Alan's "HiperSocket VSWITCH Bridge" could be an option, but multiple people (including Alan) have recommended using direct Hipersocket device definitions for each zLinux user. If I make it simple and define unique devices to each user, how do I fail over to a different device if MVS moves to another machine? I'm also going to be limited as to the number of zLinux users on any given CHPID, or is that number high enough that its not an issue? Chet Norris -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets over Guest LAN to MVS
On Monday, 07/16/2012 at 10:51 EDT, Shane G wrote: > On Tue, Jul 17th, 2012 at 6:37 AM, Alan Altmark wrote: > ... > > TCPIP is connected to two HiperSocket networks: one real (to MVS) and one > > virtual (to Linux). > > Now, hold it right there fella. I want the order number for one of those > *real* hipersockets. > (haven't we been here before ... ;-) The ephemeral energy-to-matter converter modules are already included at no additional charge. ;-) Alan Altmark Senior Managing z/VM and Linux Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets over Guest LAN to MVS
On Tue, Jul 17th, 2012 at 6:37 AM, Alan Altmark wrote: ... > TCPIP is connected to two HiperSocket networks: one real (to MVS) and one > virtual (to Linux). Now, hold it right there fella. I want the order number for one of those *real* hipersockets. (haven't we been here before ... ;-) Shane ... -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets over Guest LAN to MVS
On Monday, 07/16/2012 at 09:44 EDT, Mark Post wrote: > Alternatively, he could just attach a real HiperSocket device to the Linux > system and cut out the middle man (z/VM TCPIP) entirely. Unless there's some > value to be had by having z/VM manage the traffic, it's just unnecessary > overhead. Yes, if you don't need to enforce limits on what the guest can do in the network, then you can have it directly access the HiperSocket. And if you think, that MVS and VM might not always be in the same CEC, then you need a different topology or a different technology (HiperSocket VSWITCH Bridge). Alan Altmark Senior Managing z/VM and Linux Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets over Guest LAN to MVS
>>> On 7/16/2012 at 04:37 PM, Alan Altmark wrote: > TCPIP is connected to two HiperSocket networks: one real (to MVS) and one > virtual (to Linux). So TCPIP will need two HiperSocket interfaces, each > with it's own IP address in each of the two networks. You have the real > one (5.1.1.1), but you are missing 5.1.2.x, which is what Linux would use > as its default gateway, not 5.1.1.1. You can't ping 5.1.1.1 from Linux > since it isn't directly attached to the Linux guest and you don't have a > reachable gateway. Alternatively, he could just attach a real HiperSocket device to the Linux system and cut out the middle man (z/VM TCPIP) entirely. Unless there's some value to be had by having z/VM manage the traffic, it's just unnecessary overhead. Mark Post -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets over Guest LAN to MVS
On Monday, 07/16/2012 at 02:38 EDT, "Norris, Chet" wrote: > I'm not able to get any response back to my zlinux guest over a guest LAN I've > set up for Hipersocket connections. > > I've defined IP/ADDR 5.1.1.1 as a VM home address, 5.1.1.3 as the MVS home and > set up the Linux guest as 5.1.2.10. > A static route for 5.1.2.0/24 has been defined under MVS to point to 5.1.1.1 as > the gateway. > I've defined a guest LAN (HIPRLNX1) with type HPERS, defined special devices > 600-602 to the guest, coupled the devices to the LAN, and brought the devices > online. > I defined HSI0 to use the 600-602 devices and activated the HSI0 interface. > I've also added a network-scripts/route-hsi0 file with "default via 5.1.1.1 dev > hsi0" but to no avail. > When I ping z/VM (5.1.1.1) from zLinux (5.1.2.10) or zLinux from z/VM I get no > response. MVS (5.1.1.3) to VM (5.1.1.1) works fine, and visa versa. > Any ideas? Yes, your network diagram is messed up. Your picture is Linux [Guest LAN] TCPIP [HiperSockets] MVS TCPIP is connected to two HiperSocket networks: one real (to MVS) and one virtual (to Linux). So TCPIP will need two HiperSocket interfaces, each with it's own IP address in each of the two networks. You have the real one (5.1.1.1), but you are missing 5.1.2.x, which is what Linux would use as its default gateway, not 5.1.1.1. You can't ping 5.1.1.1 from Linux since it isn't directly attached to the Linux guest and you don't have a reachable gateway. BTW, you can use the ifconfig command on VM to get a much better display of the VM config. As an aside, I don't recommend the use of HiperSocket Guest LANs unless it is for testing purposes. Instead, use a virtual switch without an OSA. The access controls on VSWITCHes are far better than it is for Guest LANs. Alan Altmark Senior Managing z/VM and Linux Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets over Guest LAN to MVS
On Mon, Jul 16, 2012 at 04:59:49PM +, Norris, Chet wrote: > I've defined IP/ADDR 5.1.1.1 as a VM home address, 5.1.1.3 as the MVS home > and set up the Linux guest as 5.1.2.10. It would be preferable if you could use RFC 1918 allocations for your legacy IP interconnections. Just using random IP addresses does break connectivity, especially when keeping it afterwards. Of course if you argue that no internet connectivity whatsoever exists for this machine, you need to make sure that this will be the case forever. As for the specific HSI issue I cannot help you, though. Sorry. Kind regards Philipp Kern -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets
I think I found the problem. The z/os systems are defined in LCSS 0 and the zlinux is in LCSS 2. The channels are defined as shared they should be spanned. -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of McKown, John Sent: Wednesday, June 27, 2012 8:26 AM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets Another silly question from me. Is the z/Linux system using iptables to do firewalling? I don't know what z/Linux installs as a default. -- John McKown Systems Engineer IV IT Administrative Services Group HealthMarkets(r) 9151 Boulevard 26 * N. Richland Hills * TX 76010 (817) 255-3225 phone * john.mck...@healthmarkets.com * www.HealthMarkets.com Confidentiality Notice: This e-mail message may contain confidential or proprietary information. If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies of the original message. HealthMarkets(r) is the brand name for products underwritten and issued by the insurance subsidiaries of HealthMarkets, Inc. -The Chesapeake Life Insurance Company(r), Mid-West National Life Insurance Company of TennesseeSM and The MEGA Life and Health Insurance Company.SM > -Original Message- > From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of > Schwartz, Rodney (R.) > Sent: Wednesday, June 27, 2012 5:49 AM > To: LINUX-390@VM.MARIST.EDU > Subject: Re: Hipersockets > > The ping times out from both the z/os and zlinux. We have multiple > z/os systems and they can ping each other, but the zlinux. > > -Original Message- > From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of > McKown, John > Sent: Tuesday, June 26, 2012 4:36 PM > To: LINUX-390@VM.MARIST.EDU > Subject: Re: Hipersockets > > If the OP wants to look on the z/OS side, I can think of three > methods. > > z/OS console command: > > D TCPIP,,NETSTAT,DEVLINKS > > D TCPIP,,NETSTAT,ROUTE > > D TCPIP,,NETSTAT,HOME > > z/OS TSO command: > > NETSTAT DEVLINKS > > NETSTAT ROUTE > > NETSTAT HOME > > z/OS UNIX command: > > netstat -d > > netstat -r > > netstat -h > > The DEVLINKS / -d lists all the active interfaces, kind of like the > ifconfig command. But lacks the IP address. > > The ROUTE / -d lists the routes, similar to the Linux route command. > > The HOME / -h lists the IP addresses and interfaces on z/OS. > > > Can you ping the z/Linux system from z/OS? > > -- > John McKown > Systems Engineer IV > IT > > Administrative Services Group > > HealthMarkets(r) > > 9151 Boulevard 26 * N. Richland Hills * TX 76010 > (817) 255-3225 phone * > john.mck...@healthmarkets.com * www.HealthMarkets.com > > Confidentiality Notice: This e-mail message may contain confidential > or proprietary information. If you are not the intended recipient, > please contact the sender by reply e-mail and destroy all copies of > the original message. > HealthMarkets(r) is the brand name for products underwritten and > issued by the insurance subsidiaries of HealthMarkets, Inc. -The > Chesapeake Life Insurance Company(r), Mid-West National Life Insurance > Company of TennesseeSM and The MEGA Life and Health Insurance > Company.SM > > -- > For LINUX-390 subscribe / signoff / archive access instructions, send > email to lists...@vm.marist.edu with the > message: INFO LINUX-390 or visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > -- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ > > -- > For LINUX-390 subscribe / signoff / archive access instructions, send > email to lists...@vm.marist.edu with the message: INFO > LINUX-390 or visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > -- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ > > -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets
Another silly question from me. Is the z/Linux system using iptables to do firewalling? I don't know what z/Linux installs as a default. -- John McKown Systems Engineer IV IT Administrative Services Group HealthMarkets(r) 9151 Boulevard 26 * N. Richland Hills * TX 76010 (817) 255-3225 phone * john.mck...@healthmarkets.com * www.HealthMarkets.com Confidentiality Notice: This e-mail message may contain confidential or proprietary information. If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies of the original message. HealthMarkets(r) is the brand name for products underwritten and issued by the insurance subsidiaries of HealthMarkets, Inc. -The Chesapeake Life Insurance Company(r), Mid-West National Life Insurance Company of TennesseeSM and The MEGA Life and Health Insurance Company.SM > -Original Message- > From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On > Behalf Of Schwartz, Rodney (R.) > Sent: Wednesday, June 27, 2012 5:49 AM > To: LINUX-390@VM.MARIST.EDU > Subject: Re: Hipersockets > > The ping times out from both the z/os and zlinux. We have > multiple z/os systems and they can ping each other, but the zlinux. > > -Original Message- > From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On > Behalf Of McKown, John > Sent: Tuesday, June 26, 2012 4:36 PM > To: LINUX-390@VM.MARIST.EDU > Subject: Re: Hipersockets > > If the OP wants to look on the z/OS side, I can think of > three methods. > > z/OS console command: > > D TCPIP,,NETSTAT,DEVLINKS > > D TCPIP,,NETSTAT,ROUTE > > D TCPIP,,NETSTAT,HOME > > z/OS TSO command: > > NETSTAT DEVLINKS > > NETSTAT ROUTE > > NETSTAT HOME > > z/OS UNIX command: > > netstat -d > > netstat -r > > netstat -h > > The DEVLINKS / -d lists all the active interfaces, kind of > like the ifconfig command. But lacks the IP address. > > The ROUTE / -d lists the routes, similar to the Linux route command. > > The HOME / -h lists the IP addresses and interfaces on z/OS. > > > Can you ping the z/Linux system from z/OS? > > -- > John McKown > Systems Engineer IV > IT > > Administrative Services Group > > HealthMarkets(r) > > 9151 Boulevard 26 * N. Richland Hills * TX 76010 > (817) 255-3225 phone * > john.mck...@healthmarkets.com * www.HealthMarkets.com > > Confidentiality Notice: This e-mail message may contain > confidential or proprietary information. If you are not the > intended recipient, please contact the sender by reply e-mail > and destroy all copies of the original message. > HealthMarkets(r) is the brand name for products underwritten > and issued by the insurance subsidiaries of HealthMarkets, > Inc. -The Chesapeake Life Insurance Company(r), Mid-West > National Life Insurance Company of TennesseeSM and The MEGA > Life and Health Insurance Company.SM > > -- > For LINUX-390 subscribe / signoff / archive access > instructions, send email to lists...@vm.marist.edu with the > message: INFO LINUX-390 or visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > -- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ > > -- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to lists...@vm.marist.edu with the message: INFO > LINUX-390 or visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > -- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ > > -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets
Correct the address of z/os is 192.168.1.11 -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Mark Post Sent: Tuesday, June 26, 2012 4:13 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets >>> On 6/26/2012 at 02:29 PM, Marcy Cortes >>> wrote: > "ip address is 192.169.1.11" > > I hope that was a typo and you meant the address of z/OS is 192.168.1.11? One way to figure that out would be to start tcpdump on the hsi0 interface and see what kind of IP addresses show up from other systems on the subnet. I've always found that very informative in cases where I wasn't sure the network parameters I'd been given were actually right. Mark Post -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets
The ping times out from both the z/os and zlinux. We have multiple z/os systems and they can ping each other, but the zlinux. -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of McKown, John Sent: Tuesday, June 26, 2012 4:36 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets If the OP wants to look on the z/OS side, I can think of three methods. z/OS console command: D TCPIP,,NETSTAT,DEVLINKS D TCPIP,,NETSTAT,ROUTE D TCPIP,,NETSTAT,HOME z/OS TSO command: NETSTAT DEVLINKS NETSTAT ROUTE NETSTAT HOME z/OS UNIX command: netstat -d netstat -r netstat -h The DEVLINKS / -d lists all the active interfaces, kind of like the ifconfig command. But lacks the IP address. The ROUTE / -d lists the routes, similar to the Linux route command. The HOME / -h lists the IP addresses and interfaces on z/OS. Can you ping the z/Linux system from z/OS? -- John McKown Systems Engineer IV IT Administrative Services Group HealthMarkets(r) 9151 Boulevard 26 * N. Richland Hills * TX 76010 (817) 255-3225 phone * john.mck...@healthmarkets.com * www.HealthMarkets.com Confidentiality Notice: This e-mail message may contain confidential or proprietary information. If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies of the original message. HealthMarkets(r) is the brand name for products underwritten and issued by the insurance subsidiaries of HealthMarkets, Inc. -The Chesapeake Life Insurance Company(r), Mid-West National Life Insurance Company of TennesseeSM and The MEGA Life and Health Insurance Company.SM -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets
If the OP wants to look on the z/OS side, I can think of three methods. z/OS console command: D TCPIP,,NETSTAT,DEVLINKS D TCPIP,,NETSTAT,ROUTE D TCPIP,,NETSTAT,HOME z/OS TSO command: NETSTAT DEVLINKS NETSTAT ROUTE NETSTAT HOME z/OS UNIX command: netstat -d netstat -r netstat -h The DEVLINKS / -d lists all the active interfaces, kind of like the ifconfig command. But lacks the IP address. The ROUTE / -d lists the routes, similar to the Linux route command. The HOME / -h lists the IP addresses and interfaces on z/OS. Can you ping the z/Linux system from z/OS? -- John McKown Systems Engineer IV IT Administrative Services Group HealthMarkets(r) 9151 Boulevard 26 * N. Richland Hills * TX 76010 (817) 255-3225 phone * john.mck...@healthmarkets.com * www.HealthMarkets.com Confidentiality Notice: This e-mail message may contain confidential or proprietary information. If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies of the original message. HealthMarkets(r) is the brand name for products underwritten and issued by the insurance subsidiaries of HealthMarkets, Inc. -The Chesapeake Life Insurance Company(r), Mid-West National Life Insurance Company of TennesseeSM and The MEGA Life and Health Insurance Company.SM -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets
>>> On 6/26/2012 at 02:29 PM, Marcy Cortes >>> wrote: > "ip address is 192.169.1.11" > > I hope that was a typo and you meant the address of z/OS is 192.168.1.11? One way to figure that out would be to start tcpdump on the hsi0 interface and see what kind of IP addresses show up from other systems on the subnet. I've always found that very informative in cases where I wasn't sure the network parameters I'd been given were actually right. Mark Post -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets
"ip address is 192.169.1.11" I hope that was a typo and you meant the address of z/OS is 192.168.1.11? Marcy -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Schwartz, Rodney (R.) Sent: Tuesday, June 26, 2012 11:11 AM To: LINUX-390@VM.MARIST.EDU Subject: [LINUX-390] Hipersockets I am trying to get Hipersockets working between z/os and zlinux system (sles 10 SP2 running under z/vm 6.2). I have a dedicated IOD address attached: OSA FB00 ATTACHED TO LINUX02 7000 DEVTYPE HIPER CHPID FB IQD OSA FB01 ATTACHED TO LINUX02 7001 DEVTYPE HIPER CHPID FB IQD OSA FB02 ATTACHED TO LINUX02 7002 DEVTYPE HIPER CHPID FB IQD lscss | grep 1732 0.0.7000 0.0.0003 1732/05 1731/05 yes 80 80 FF FB00 0.0.7001 0.0.0004 1732/05 1731/05 yes 80 80 FF FB00 0.0.7002 0.0.0005 1732/05 1731/05 yes 80 80 FF FB00 LINUX02:~ # ifconfig eth0 Link encap:Ethernet HWaddr 02:00:04:00:00:0A inet addr:19.59.28.7 Bcast:19.59.29.255 Mask:255.255.254.0 UP BROADCAST RUNNING MULTICAST MTU:1492 Metric:1 RX packets:1217 errors:0 dropped:0 overruns:0 frame:0 TX packets:997 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:106554 (104.0 Kb) TX bytes:149236 (145.7 Kb) hsi0 Link encap:Ethernet HWaddr 06:00:FB:24:00:02 inet addr:192.168.1.25 Bcast:192.168.1.255 Mask:255.255.255.0 UP BROADCAST RUNNING NOARP MULTICAST MTU:57344 Metric:1 RX packets:0 errors:0 dropped:0 overruns:0 frame:0 TX packets:20 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:0 (0.0 b) TX bytes:1680 (1.6 Kb) loLink encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 UP LOOPBACK RUNNING MTU:16436 Metric:1 RX packets:23 errors:0 dropped:0 overruns:0 frame:0 TX packets:23 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:1766 (1.7 Kb) TX bytes:1766 (1.7 Kb) LINUX02:~ # routes Kernel IP routing table Destination Gateway Genmask Flags Metric RefUse Iface 192.168.1.0 * 255.255.255.0 U 0 00 hsi0 19.59.28.0 * 255.255.254.0 U 0 00 eth0 link-local * 255.255.0.0 U 0 00 eth0 loopback* 255.0.0.0 U 0 00 lo default fmcrsd1v228sb.n 0.0.0.0 UG0 00 eth0 The z/os hipesocket address is on the same chpid. The ip address is 192.169.1.11 We have multiple z/os systems sharing the same chpid. They can all talk through the hipersocket, But the zlinux system just times out when I ping z/os I know I am missing something, but I do not know what. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets don't come up after HW upgrade
That worked - the error msg. went away, and it shaved about 20 seconds off the boot time. This was a new clone of SLES11sp1. So the file must have been created by yast or something else I did while trying to fix the problem. Since you're paying taxes in Germany, I suppose you're doing your part to fix the other crisis and we appreciate that, too. Best Regards Roger Evans, Autodata Norge A/S http://www.autodata.no On Tue, 2011-11-08 at 13:18 +0100, Ursula Braun wrote: > Roger, > > good news. I am working in the networking team of Linux on System z > development; thus qeth-related problems are my daily business. But > please do not ask me financial questions ;-) > > For your remaining problem: Have you upgraded your system from SLES10? > In this case you may still have old configuration > files /etc/sysconfig/hardware/hwcfg-hsi-bus-ccw-0.0.7000 > or /etc/sysconfig/network/ifcfg-hsi-bus-ccw-0.0.7000. Those are no > longer needed for SLES11 and can be removed. > > Ursula > > On Tue, 2011-11-08 at 12:58 +0100, Roger Evans wrote: > > Thank you, Ursula > > That solved the problem. I can now ping and ftp between my layer3 > > hipersockets. I still get the "Waiting for mandatory devices. > > hsi-bus-ccw-0.0.7000 No interface found " msg. when booting > > but now that it works, I'll just stop watching. > > > > You're pretty amazing. Think you could fix the Greek deficit crisis, > > too? > > > > > > Best Regards > > Roger Evans, > > Autodata Norge A/S > > http://www.autodata.no > > > > -- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > -- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets don't come up after HW upgrade
Roger, good news. I am working in the networking team of Linux on System z development; thus qeth-related problems are my daily business. But please do not ask me financial questions ;-) For your remaining problem: Have you upgraded your system from SLES10? In this case you may still have old configuration files /etc/sysconfig/hardware/hwcfg-hsi-bus-ccw-0.0.7000 or /etc/sysconfig/network/ifcfg-hsi-bus-ccw-0.0.7000. Those are no longer needed for SLES11 and can be removed. Ursula On Tue, 2011-11-08 at 12:58 +0100, Roger Evans wrote: > Thank you, Ursula > That solved the problem. I can now ping and ftp between my layer3 > hipersockets. I still get the "Waiting for mandatory devices. > hsi-bus-ccw-0.0.7000 No interface found " msg. when booting > but now that it works, I'll just stop watching. > > You're pretty amazing. Think you could fix the Greek deficit crisis, > too? > > > Best Regards > Roger Evans, > Autodata Norge A/S > http://www.autodata.no > -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets don't come up after HW upgrade
Thank you, Ursula That solved the problem. I can now ping and ftp between my layer3 hipersockets. I still get the "Waiting for mandatory devices. hsi-bus-ccw-0.0.7000 No interface found " msg. when booting but now that it works, I'll just stop watching. You're pretty amazing. Think you could fix the Greek deficit crisis, too? Best Regards Roger Evans, Autodata Norge A/S http://www.autodata.no On Tue, 2011-11-08 at 12:22 +0100, Ursula Braun wrote: > Roger, > > get rid of the LLADDR='00:00:00:00:00:00' definition. ifup does no > longer work for a zero MAC-address. HiperSockets define a MAC-address; > during initialization Linux determines this MAC-address and uses it, if > LLADDR is not defined. > > Ursula > > On Tue, 2011-11-08 at 10:47 +0100, Roger Evans wrote: > > Forgot to send the info you requested: > > > > DPRODDB2:/etc/sysconfig/network # cat ifcfg-hsi0 > > BOOTPROTO='static' > > BROADCAST='' > > ETHTOOL_OPTIONS='' > > IPADDR='10.5.2.11/16' > > LLADDR='00:00:00:00:00:00' > > MTU='' > > NAME='Hipersocket (0.0.7000)' > > NETWORK='' > > REMOTE_IPADDR='' > > STARTMODE='manual' > > USERCONTROL='no' > > > > Roger > > > > -- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > -- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets don't come up after HW upgrade
Roger, when configuring a network interface with yast, a udev-rule is created for the device containing qeth attribute definitions, among them attribute layer2. Look into /etc/udev/rules.d and check a rule starting with 51-...7000... . I assume this rule contains a line with layer2=1. Change this into layer2=0 and the device should come up with layer3 after reboot. Ursula On Tue, 2011-11-08 at 10:45 +0100, Roger Evans wrote: > THanks for that tip. I can't remember having chosen level2, but I see > that my other (older) interfaces are not level2, so, yes, I would like > to switch to level3. > > I used the commands (from "Device Drivers, Features...") to to take the > device offline and to make it not level2 (level2=9); After that, lsqeth > showed it as level2=0. > But after rebooting, it reverted to level2=1. > > I noticed the following message on VM when booting: > --- > ..skippedWaiting for mandatory devices: hsi-bus-ccw-0.0.7000 __NSC__ > 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 > 3 2 1 0 > > hsi-bus-ccw-0.0.7000No interface found > > > Any idea what I'm missing? > > Roger > -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets don't come up after HW upgrade
Roger, get rid of the LLADDR='00:00:00:00:00:00' definition. ifup does no longer work for a zero MAC-address. HiperSockets define a MAC-address; during initialization Linux determines this MAC-address and uses it, if LLADDR is not defined. Ursula On Tue, 2011-11-08 at 10:47 +0100, Roger Evans wrote: > Forgot to send the info you requested: > > DPRODDB2:/etc/sysconfig/network # cat ifcfg-hsi0 > BOOTPROTO='static' > BROADCAST='' > ETHTOOL_OPTIONS='' > IPADDR='10.5.2.11/16' > LLADDR='00:00:00:00:00:00' > MTU='' > NAME='Hipersocket (0.0.7000)' > NETWORK='' > REMOTE_IPADDR='' > STARTMODE='manual' > USERCONTROL='no' > > Roger > -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets don't come up after HW upgrade
Forgot to send the info you requested: DPRODDB2:/etc/sysconfig/network # cat ifcfg-hsi0 BOOTPROTO='static' BROADCAST='' ETHTOOL_OPTIONS='' IPADDR='10.5.2.11/16' LLADDR='00:00:00:00:00:00' MTU='' NAME='Hipersocket (0.0.7000)' NETWORK='' REMOTE_IPADDR='' STARTMODE='manual' USERCONTROL='no' Roger On Tue, 2011-11-08 at 09:37 +0100, Ursula Braun wrote: > Hi Roger, > > your boot.messages do not show any message refering to 0.0.7000 or hsi0. > Your hsi0-device is in state SOFTSETUP, which means it has been > activated successfully; just the ifup step fails. The configuration file > for ifup is /etc/sysconfig/network/ifcfg-hsi0. Can you check the > definitions of this file? > > By the way, you have chosen layer2=1 for your hsi0-device. This is ok, > but it allows communication to other layer2 HiperSockets participants > only, no layer3 HiperSockets participants. > > Regards, Ursula Braun, IBM Germany > > On Tue, 2011-11-08 at 08:39 +0100, Roger Evans wrote: > > Hi, Ursula > > > > uname -a > > Linux DPRODDB2 2.6.32.46-0.3-default #1 SMP 2011-09-29 17:49:31 +0200 > > s390x s390x s390x GNU/Linux > > > > lsqeth > > ... [other devices] > > Device name : hsi0 > > - > > card_type : HiperSockets > > cdev0 : 0.0.7000 > > cdev1 : 0.0.7001 > > cdev2 : 0.0.7002 > > chpid : F1 > > online : 1 > > portname: no portname required > > portno : 0 > > state : SOFTSETUP > > priority_queueing : always queue 2 > > buffer_count: 16 > > layer2 : 1 > > isolation : none > > > > I am attaching the results of the dmesg command > > > > > > Med vennlig hilsen > > > > Roger Evans > > -- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > -- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets don't come up after HW upgrade
THanks for that tip. I can't remember having chosen level2, but I see that my other (older) interfaces are not level2, so, yes, I would like to switch to level3. I used the commands (from "Device Drivers, Features...") to to take the device offline and to make it not level2 (level2=9); After that, lsqeth showed it as level2=0. But after rebooting, it reverted to level2=1. I noticed the following message on VM when booting: --- ..skippedWaiting for mandatory devices: hsi-bus-ccw-0.0.7000 __NSC__ 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 hsi-bus-ccw-0.0.7000No interface found Any idea what I'm missing? Roger On Tue, 2011-11-08 at 09:37 +0100, Ursula Braun wrote: > Hi Roger, > > your boot.messages do not show any message refering to 0.0.7000 or hsi0. > Your hsi0-device is in state SOFTSETUP, which means it has been > activated successfully; just the ifup step fails. The configuration file > for ifup is /etc/sysconfig/network/ifcfg-hsi0. Can you check the > definitions of this file? > > By the way, you have chosen layer2=1 for your hsi0-device. This is ok, > but it allows communication to other layer2 HiperSockets participants > only, no layer3 HiperSockets participants. > > Regards, Ursula Braun, IBM Germany > > On Tue, 2011-11-08 at 08:39 +0100, Roger Evans wrote: > > Hi, Ursula > > > > uname -a > > Linux DPRODDB2 2.6.32.46-0.3-default #1 SMP 2011-09-29 17:49:31 +0200 > > s390x s390x s390x GNU/Linux > > > > lsqeth > > ... [other devices] > > Device name : hsi0 > > - > > card_type : HiperSockets > > cdev0 : 0.0.7000 > > cdev1 : 0.0.7001 > > cdev2 : 0.0.7002 > > chpid : F1 > > online : 1 > > portname: no portname required > > portno : 0 > > state : SOFTSETUP > > priority_queueing : always queue 2 > > buffer_count: 16 > > layer2 : 1 > > isolation : none > > > > I am attaching the results of the dmesg command > > > > > > Med vennlig hilsen > > > > Roger Evans > > -- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > -- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets don't come up after HW upgrade
Hi Roger, your boot.messages do not show any message refering to 0.0.7000 or hsi0. Your hsi0-device is in state SOFTSETUP, which means it has been activated successfully; just the ifup step fails. The configuration file for ifup is /etc/sysconfig/network/ifcfg-hsi0. Can you check the definitions of this file? By the way, you have chosen layer2=1 for your hsi0-device. This is ok, but it allows communication to other layer2 HiperSockets participants only, no layer3 HiperSockets participants. Regards, Ursula Braun, IBM Germany On Tue, 2011-11-08 at 08:39 +0100, Roger Evans wrote: > Hi, Ursula > > uname -a > Linux DPRODDB2 2.6.32.46-0.3-default #1 SMP 2011-09-29 17:49:31 +0200 > s390x s390x s390x GNU/Linux > > lsqeth > ... [other devices] > Device name : hsi0 > - > card_type : HiperSockets > cdev0 : 0.0.7000 > cdev1 : 0.0.7001 > cdev2 : 0.0.7002 > chpid : F1 > online : 1 > portname: no portname required > portno : 0 > state : SOFTSETUP > priority_queueing : always queue 2 > buffer_count: 16 > layer2 : 1 > isolation : none > > I am attaching the results of the dmesg command > > > Med vennlig hilsen > > Roger Evans -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets don't come up after HW upgrade
Hi, Ursula uname -a Linux DPRODDB2 2.6.32.46-0.3-default #1 SMP 2011-09-29 17:49:31 +0200 s390x s390x s390x GNU/Linux lsqeth ... [other devices] Device name : hsi0 - card_type : HiperSockets cdev0 : 0.0.7000 cdev1 : 0.0.7001 cdev2 : 0.0.7002 chpid : F1 online : 1 portname: no portname required portno : 0 state : SOFTSETUP priority_queueing : always queue 2 buffer_count: 16 layer2 : 1 isolation : none I am attaching the results of the dmesg command Med vennlig hilsen Roger Evans On Mon, 2011-11-07 at 16:53 +0100, Ursula Braun wrote: > Roger, > > the minimum SLES11 SP1 kernel level should be 2.6.32.29-0.3.1. If you > still have problems with a kernel level greater or equal to this one, > something else is wrong. Which qeth-related messages show up in dmesg? > What does "lsqeth" return? > > Regards, Ursula Braun, IBM Germany > > On Mon, 2011-11-07 at 16:04 +0100, Roger Evans wrote: > > Thank you Ursula > > > > I don't seem to have the permissions to read those bugzilla entries, or > > maybe I'm looking in the wrong Novell Bugzilla. > > > > Upgrading from sp3 to sp4 solved the problem for SLES10. > > > > I upgraded my SLES11sp1 with all the upgrades SuSE provides, but still > > get the following msg: > > -- > > DPRODDB2:~ # ifup hsi0 > > hsi0 name: Hipersocket (0.0.7000) > > RTNETLINK answers: Unknown error 18446744073709486085 > > Cannot enable interface hsi0. > > interface hsi0 is not up > > - > > > > I will go through the steps in the hipersocket configuration > > instructions again. Maybe it's not enough to define it in YAST.. > > > > -- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > -- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ Initializing cgroup subsys cpuset Initializing cgroup subsys cpu Linux version 2.6.32.46-0.3-default (geeko@buildhost) (gcc version 4.3.4 [gcc-4_3-branch revision 152973] (SUSE Linux) ) #1 SMP 2011-09-29 17:49:31 +0200 setup.1a06a7: Linux is running as a z/VM guest operating system in 64-bit mode Zone PFN ranges: DMA 0x -> 0x0008 Normal 0x0008 -> 0x0010 Movable zone start PFN for each node early_node_map[1] active PFN ranges 0: 0x -> 0x0010 On node 0 totalpages: 1048576 DMA zone: 7168 pages used for memmap DMA zone: 0 pages reserved DMA zone: 517120 pages, LIFO batch:31 Normal zone: 7168 pages used for memmap Normal zone: 517120 pages, LIFO batch:31 PERCPU: Embedded 11 pages/cpu @8408 s12544 r8192 d24320 u65536 pcpu-alloc: s12544 r8192 d24320 u65536 alloc=16*4096 pcpu-alloc: [0] 00 [0] 01 [0] 02 [0] 03 [0] 04 [0] 05 [0] 06 [0] 07 pcpu-alloc: [0] 08 [0] 09 [0] 10 [0] 11 [0] 12 [0] 13 [0] 14 [0] 15 pcpu-alloc: [0] 16 [0] 17 [0] 18 [0] 19 [0] 20 [0] 21 [0] 22 [0] 23 pcpu-alloc: [0] 24 [0] 25 [0] 26 [0] 27 [0] 28 [0] 29 [0] 30 [0] 31 pcpu-alloc: [0] 32 [0] 33 [0] 34 [0] 35 [0] 36 [0] 37 [0] 38 [0] 39 pcpu-alloc: [0] 40 [0] 41 [0] 42 [0] 43 [0] 44 [0] 45 [0] 46 [0] 47 pcpu-alloc: [0] 48 [0] 49 [0] 50 [0] 51 [0] 52 [0] 53 [0] 54 [0] 55 pcpu-alloc: [0] 56 [0] 57 [0] 58 [0] 59 [0] 60 [0] 61 [0] 62 [0] 63 Built 1 zonelists in Zone order, mobility grouping on. Total pages: 1034240 Kernel command line: root=/dev/disk/by-path/ccw-0.0.0100-part1 TERM=dumb BOOT_IMAGE=0 PID hash table entries: 4096 (order: 3, 32768 bytes) Dentry cache hash table entries: 524288 (order: 10, 4194304 bytes) Inode-cache hash table entries: 262144 (order: 9, 2097152 bytes) Memory: 4105820k/4194304k available (4398k kernel code, 0k reserved, 2099k data, 228k init) Write protected kernel read-only data: 0x10 - 0x5f Hierarchical RCU implementation. console [ttyS0] enabled allocated 41943040 bytes of page_cgroup please try 'cgroup_disable=memory' option if you
Re: hipersockets don't come up after HW upgrade
Roger, the minimum SLES11 SP1 kernel level should be 2.6.32.29-0.3.1. If you still have problems with a kernel level greater or equal to this one, something else is wrong. Which qeth-related messages show up in dmesg? What does "lsqeth" return? Regards, Ursula Braun, IBM Germany On Mon, 2011-11-07 at 16:04 +0100, Roger Evans wrote: > Thank you Ursula > > I don't seem to have the permissions to read those bugzilla entries, or > maybe I'm looking in the wrong Novell Bugzilla. > > Upgrading from sp3 to sp4 solved the problem for SLES10. > > I upgraded my SLES11sp1 with all the upgrades SuSE provides, but still > get the following msg: > -- > DPRODDB2:~ # ifup hsi0 > hsi0 name: Hipersocket (0.0.7000) > RTNETLINK answers: Unknown error 18446744073709486085 > Cannot enable interface hsi0. > interface hsi0 is not up > - > > I will go through the steps in the hipersocket configuration > instructions again. Maybe it's not enough to define it in YAST.. > -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets don't come up after HW upgrade
Thank you Ursula I don't seem to have the permissions to read those bugzilla entries, or maybe I'm looking in the wrong Novell Bugzilla. Upgrading from sp3 to sp4 solved the problem for SLES10. I upgraded my SLES11sp1 with all the upgrades SuSE provides, but still get the following msg: -- DPRODDB2:~ # ifup hsi0 hsi0 name: Hipersocket (0.0.7000) RTNETLINK answers: Unknown error 18446744073709486085 Cannot enable interface hsi0. interface hsi0 is not up - I will go through the steps in the hipersocket configuration instructions again. Maybe it's not enough to define it in YAST.. Regards/Med vennlig hilsen Roger Evans Autodata Norge A/S On Mon, 2011-11-07 at 10:06 +0100, Ursula Braun wrote: > On Fri, 2011-11-04 at 15:18 +0200, Roger Evans wrote: > > Hi Sue > > > > Didn't seem to help. When booting with SLES10v2 I see the following > > when watching the boot from VM: > > - > > eth: IPv6 not supported on hsi0 > > qeth: Broadcast enabled > > qeth: Using SW checksumming on hsi0. > > qeth: Outbound TSO not supported on hsi0 > > operand exception: 0015 ¬Æ1| > > CPU: 0 Not tainted > > Process hwup-qeth (pid: 709, task: 0f400048, ksp: > > 0eee3960) > > Krnl PSW : 070400018000 1098cefa (do_QDIO+0x452/0x2a9c > > ¬qdio|) > > Krnl GPRS: 0001 00010006 8000 > > 8000 > > 00010006 000f > > 0e931000 0e4cf000 0e931000 > > 1097f000 10994960 0eee3ba8 0eee3aa8 > > Krnl Code: b2 22 00 30 88 30 00 1c 12 33 a7 84 00 0c bf 4f 92 50 a7 84 > > Call Trace: > > (¬<10b03a68>| qeth_dbf_setup+0x0/0xfffe7ad0 ¬qeth|) > > ¬<10adeafc>| __qeth_set_online+0x2780/0x2de4 ¬qeth| > > ¬<1081b544>| ccwgroup_online_store+0x10c/0x1fc ¬ccwgroup| > > ¬<002c43a0>| sysfs_write_file+0x120/0x190 > > ¬<002250f8>| sys_write+0x188/0x38c > > ¬<00115d10>| sysc_noemu+0x10/0x16 > > ¬<021c8738>| 0x21c8738 > > ..done > > Loading required kernel modules > > ¬1A..done<6>device-mapper: ioctl: 4.7.0-ioctl (2006-06-24) initialised: > > dm-devel > > ... > > -- > > > > I get this regardless of whether QIOASSIST is ON or OFF. > > > > Regards/ > > Med vennlig hilsen > > > > > > Roger Evans > > > Roger, > > the HiperSockets issue after HW upgrade requires a Linux upgrade: > - see Novell bugzilla 659101 for SLES11 SP1 > - see Novell bugzilla 662984 for SLES10 SP4 > - RHEL 5.7 > - RHEL 6.1 > > Description: qdio: use proper QEBSM operand for SIGA-R and SIGA-S > Symptom: Operand exception leading to kernel panic. > Problem: Wrong SIGA operand on QEBSM enabled qdio devices. > Solution:Use proper SIGA operands in QEBSM mode. > > Switching QIOASSIST OFF for HiperSockets Devices should circumvent the > problem. > > Regards, Ursula Braun, IBM Germany > > > -- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > -- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets don't come up after HW upgrade
On Fri, 2011-11-04 at 15:18 +0200, Roger Evans wrote: > Hi Sue > > Didn't seem to help. When booting with SLES10v2 I see the following > when watching the boot from VM: > - > eth: IPv6 not supported on hsi0 > qeth: Broadcast enabled > qeth: Using SW checksumming on hsi0. > qeth: Outbound TSO not supported on hsi0 > operand exception: 0015 ¬Æ1| > CPU: 0 Not tainted > Process hwup-qeth (pid: 709, task: 0f400048, ksp: > 0eee3960) > Krnl PSW : 070400018000 1098cefa (do_QDIO+0x452/0x2a9c > ¬qdio|) > Krnl GPRS: 0001 00010006 8000 > 8000 > 00010006 000f > 0e931000 0e4cf000 0e931000 > 1097f000 10994960 0eee3ba8 0eee3aa8 > Krnl Code: b2 22 00 30 88 30 00 1c 12 33 a7 84 00 0c bf 4f 92 50 a7 84 > Call Trace: > (¬<10b03a68>| qeth_dbf_setup+0x0/0xfffe7ad0 ¬qeth|) > ¬<10adeafc>| __qeth_set_online+0x2780/0x2de4 ¬qeth| > ¬<1081b544>| ccwgroup_online_store+0x10c/0x1fc ¬ccwgroup| > ¬<002c43a0>| sysfs_write_file+0x120/0x190 > ¬<002250f8>| sys_write+0x188/0x38c > ¬<00115d10>| sysc_noemu+0x10/0x16 > ¬<021c8738>| 0x21c8738 > ..done > Loading required kernel modules > ¬1A..done<6>device-mapper: ioctl: 4.7.0-ioctl (2006-06-24) initialised: > dm-devel > ... > -- > > I get this regardless of whether QIOASSIST is ON or OFF. > > Regards/ > Med vennlig hilsen > > > Roger Evans > Roger, the HiperSockets issue after HW upgrade requires a Linux upgrade: - see Novell bugzilla 659101 for SLES11 SP1 - see Novell bugzilla 662984 for SLES10 SP4 - RHEL 5.7 - RHEL 6.1 Description: qdio: use proper QEBSM operand for SIGA-R and SIGA-S Symptom: Operand exception leading to kernel panic. Problem: Wrong SIGA operand on QEBSM enabled qdio devices. Solution:Use proper SIGA operands in QEBSM mode. Switching QIOASSIST OFF for HiperSockets Devices should circumvent the problem. Regards, Ursula Braun, IBM Germany -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets don't come up after HW upgrade
Display the kernel storage at 10b03a60.8 - in there is the instruction giving the exception. The b2220030 is the IPM used to retrieve the condition code from the previous instruction so it can be placed in a C variable. On 11/4/11 9:18 AM, "Roger Evans" wrote: > Hi Sue > > Didn't seem to help. When booting with SLES10v2 I see the following > when watching the boot from VM: > -- > -- > - > eth: IPv6 not supported on hsi0 > qeth: Broadcast enabled > qeth: Using SW checksumming on hsi0. > qeth: Outbound TSO not supported on hsi0 > operand exception: 0015 ¬Æ1| > CPU: 0 Not tainted > Process hwup-qeth (pid: 709, task: 0f400048, ksp: > 0eee3960) > Krnl PSW : 070400018000 1098cefa (do_QDIO+0x452/0x2a9c > ¬qdio|) > Krnl GPRS: 0001 00010006 8000 > 8000 > 00010006 000f > 0e931000 0e4cf000 0e931000 > 1097f000 10994960 0eee3ba8 0eee3aa8 > Krnl Code: b2 22 00 30 88 30 00 1c 12 33 a7 84 00 0c bf 4f 92 50 a7 84 > Call Trace: > (¬<10b03a68>| qeth_dbf_setup+0x0/0xfffe7ad0 ¬qeth|) > ¬<10adeafc>| __qeth_set_online+0x2780/0x2de4 ¬qeth| > ¬<1081b544>| ccwgroup_online_store+0x10c/0x1fc ¬ccwgroup| > ¬<002c43a0>| sysfs_write_file+0x120/0x190 > ¬<002250f8>| sys_write+0x188/0x38c > ¬<00115d10>| sysc_noemu+0x10/0x16 > ¬<021c8738>| 0x21c8738 > ..done > Loading required kernel modules > ¬1A..done<6>device-mapper: ioctl: 4.7.0-ioctl (2006-06-24) initialised: > dm-devel -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets don't come up after HW upgrade
Hi Sue Didn't seem to help. When booting with SLES10v2 I see the following when watching the boot from VM: - eth: IPv6 not supported on hsi0 qeth: Broadcast enabled qeth: Using SW checksumming on hsi0. qeth: Outbound TSO not supported on hsi0 operand exception: 0015 ¬Æ1| CPU: 0 Not tainted Process hwup-qeth (pid: 709, task: 0f400048, ksp: 0eee3960) Krnl PSW : 070400018000 1098cefa (do_QDIO+0x452/0x2a9c ¬qdio|) Krnl GPRS: 0001 00010006 8000 8000 00010006 000f 0e931000 0e4cf000 0e931000 1097f000 10994960 0eee3ba8 0eee3aa8 Krnl Code: b2 22 00 30 88 30 00 1c 12 33 a7 84 00 0c bf 4f 92 50 a7 84 Call Trace: (¬<10b03a68>| qeth_dbf_setup+0x0/0xfffe7ad0 ¬qeth|) ¬<10adeafc>| __qeth_set_online+0x2780/0x2de4 ¬qeth| ¬<1081b544>| ccwgroup_online_store+0x10c/0x1fc ¬ccwgroup| ¬<002c43a0>| sysfs_write_file+0x120/0x190 ¬<002250f8>| sys_write+0x188/0x38c ¬<00115d10>| sysc_noemu+0x10/0x16 ¬<021c8738>| 0x21c8738 ..done Loading required kernel modules ¬1A..done<6>device-mapper: ioctl: 4.7.0-ioctl (2006-06-24) initialised: dm-devel ... -- I get this regardless of whether QIOASSIST is ON or OFF. Regards/ Med vennlig hilsen Roger Evans On Fri, 2011-11-04 at 08:07 -0400, Susan M. Farrell (620-3538) wrote: > Can you try turning off QIOASSIST for the guest and see > if that makes a difference? > > Regards, > Sue Farrell > > -- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > -- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets don't come up after HW upgrade
Can you try turning off QIOASSIST for the guest and see if that makes a difference? Regards, Sue Farrell -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets don't come up after HW upgrade
> What version of z/VM is this? z/VM Version 5 Release 4.0, Service Level 1101 (64-bit), > Do you have all the Linux related maintenance installed for it? Hard to tell. Is there one place we can go to find out what is needed? We have installed all the APARS that IBM told us were necessary to run VM on z114. Roger Evans Autodata Norge A/S Best Regards/Med vennlig hilsen Roger Evans Autodata Norge A/S On Tue, 2011-11-01 at 10:41 -0600, Mark Post wrote: > >>> On 11/1/2011 at 08:12 AM, Roger Evans wrote: > > SuSE 10.2 > > uname returns: kernel 2.6.16.60-0.21-default > > That's the SLES10 SP2 GA kernel, which is 2.5 years old. I think it might be > time to apply some maintenance. :( > > > I'm also trying to configure hsi0 with YAST on SuSE 11sp1 > > When I try to bring the interface up with ifup, I get: > > > > > > -- > > #ifup hsi0 > > hsi0 name: Hipersocket (0.0.7000) > > RTNETLINK answers: Unknown error 18446744073709486085 > > Cannot enable interface hsi0. > > > > -- > > What version of z/VM is this? Do you have all the Linux related maintenance > installed for it? > > > Mark Post > > -- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > -- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets don't come up after HW upgrade
>>> On 11/1/2011 at 08:12 AM, Roger Evans wrote: > SuSE 10.2 > uname returns: kernel 2.6.16.60-0.21-default That's the SLES10 SP2 GA kernel, which is 2.5 years old. I think it might be time to apply some maintenance. :( > I'm also trying to configure hsi0 with YAST on SuSE 11sp1 > When I try to bring the interface up with ifup, I get: > > > -- > #ifup hsi0 > hsi0 name: Hipersocket (0.0.7000) > RTNETLINK answers: Unknown error 18446744073709486085 > Cannot enable interface hsi0. > > -- What version of z/VM is this? Do you have all the Linux related maintenance installed for it? Mark Post -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets don't come up after HW upgrade
SuSE 10.2 uname returns: kernel 2.6.16.60-0.21-default I'm also trying to configure hsi0 with YAST on SuSE 11sp1 When I try to bring the interface up with ifup, I get: -- #ifup hsi0 hsi0 name: Hipersocket (0.0.7000) RTNETLINK answers: Unknown error 18446744073709486085 Cannot enable interface hsi0. -- Med vennlig hilsen/regards Roger Evans On Mon, 2011-10-31 at 10:03 -0600, Mark Post wrote: > >>> On 10/31/2011 at 06:00 AM, Roger Evans wrote: > > Good morning, Mailing list > > We upgraded from a z9 to a z114 during the weekend, Most things seem > > to be working fine, but none of my linux hipersocket interfaces come > > up. > > > -snip- > > None of our VM or linux definitions have changed. > > > > The console on VSE shows the hipersocket as up. > > > > Any ideas? > > What distribution? What version and maintenance level? There was a bug in > udev that required both a kernel update and udev update to correct. > > > Mark Post > > -- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > -- > For more information on Linux on System z, visit > http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: hipersockets don't come up after HW upgrade
>>> On 10/31/2011 at 06:00 AM, Roger Evans wrote: > Good morning, Mailing list > We upgraded from a z9 to a z114 during the weekend, Most things seem > to be working fine, but none of my linux hipersocket interfaces come > up. > -snip- > None of our VM or linux definitions have changed. > > The console on VSE shows the hipersocket as up. > > Any ideas? What distribution? What version and maintenance level? There was a bug in udev that required both a kernel update and udev update to correct. Mark Post -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets Not working z/Linux to z/VM & z/OS
On Tuesday, 02/01/2011 at 06:06 EST, Mark Post wrote: > What do the US and UH flags mean? What's the reason for the host route to > 10.90.3.20? U = Up (if interface is down, flag not set, route won't be used) S = Static (i.e. human-induced config somewhere) H = Host route Alan Altmark z/VM and Linux on System z Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 alan_altm...@us.ibm.com IBM Endicott -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets Not working z/Linux to z/VM & z/OS
>>> On 1/31/2011 at 01:49 PM, Kyle Stewart >>> wrote: > Mark, > > Here is what we have: > > The hsi0 is a real hipersocket > > Linux netstat > > [z034876@UTLZ0002 ~]$ netstat -r > Kernel IP routing table > Destination Gateway Genmask Flags MSS Window irtt > Iface > > 10.90.3.0 * 255.255.255.0 U 0 0 0 > hsi0 > > 10.90.30.0 * 255.255.255.0 U 0 0 0 > eth0 > > 169.254.0.0 * 255.255.0.0 U 0 0 0 > hsi0 > > default 10.90.30.1 0.0.0.0 UG0 0 0 > eth0 This looks fine so far. -snip- > z/VM netstat gate > > Known IPv4 gateways: > > Subnet Address Subnet Mask FirstHopFlgs PktSz Metric Link > -- --- - -- -- > Default10.90.1.1 UG 1500 11 OSD2 > Default10.90.1.1 UG 1500 11 OSD0 > 10.90.1.0 255.255.255.0U1500 10 OSD0 > 10.90.1.0 255.255.255.0U1500 10 OSD2 > 10.90.3.0 255.255.255.0US 16384 HIPERLFA > 10.90.11.0 255.255.255.010.90.1.12 UG 1500 20 OSD2 I don't know what effect the lack of a metric on your 10.90.3.0 route will have, but the fact that this is the network you're having problems with makes me wonder if it's related. -snip- > z/VM netstat home > > netstat home > VM TCP/IP Netstat Level 610 TCP/IP Server Name: TCPIP > > IPv4 Home address entries: > > Address Subnet Mask Link VSWITCH > --- --- ----- > 10.90.25.1 255.255.255.0VC1LVIPA > 10.90.1.14 255.255.255.0OSD2 > 10.90.1.13 255.255.255.0OSD0 > 10.90.3.60 255.255.255.0HIPERLFA I note that your z/VM OSAs are not on the same subnet as the VSWITCH your Linux systems are on. While not necessarily a problem, it's something that may have an impact on your firewall rules, if you're running any. > Test z/OS LPAR: -snip- > EZA0611I The following IP addresses correspond to TCP Host Name: ZSJES2 > EZA0612I 10.90.21.1 > > EZA0614I The following IP addresses are the HOME IP addresses defined in > PROFILE.TCPIP: > > EZA0615I 10.90.21.1 > EZA0615I 10.90.1.19 > EZA0615I 10.90.1.20 > EZA0615I 10.90.21.10 > EZA0615I 10.90.21.90 > EZA0615I 10.90.3.20 > EZA0615I 127.0.0.1 Same note here about z/OS being on a different subnet than the VSWITCH. > EZA0618I All IP addresses for ZSJES2 are in the HOME list! > EZA0622I Hometest was successful - all Tests Passed! > > > NETSTAT ROUTE > MVS TCP/IP NETSTAT CS V1R11 TCPIP Name: TCPIP 18:41:28 > DestinationGateway FlagsRefcnt Interface > ------ --- - > Default10.90.1.1 UGO 05 OSD0 > Default10.90.1.1 UGO 00 OSD2 -snip- > 10.90.3.0/24 0.0.0.0 US 00 HYPERLFA > 10.90.3.20/32 0.0.0.0 UH 00 HYPERLFA What do the US and UH flags mean? What's the reason for the host route to 10.90.3.20? As Alan mentioned, it might be interesting to see what tcpdump running on both hsi0 and eth0 shows when packets arrive and leave for the pings from z/VM and z/OS. Mark Post -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets Not working z/Linux to z/VM & z/OS
Kyle, please try to add line ARP=no to your /etc/sysconfig/network-scripts/ifcfg-hsi0 configuration file. Regards, Ursula -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets Not working z/Linux to z/VM & z/OS
On Monday, 01/31/2011 at 05:22 EST, Kyle Stewart wrote: > > We have been struggling trying to get our z/VM & z/OS systems to talk via > > hipersockets to our RHEL5.3 Linux (under z/VM 6.1). We can ping out from > > Linux to VM or to z/OS, but when we ping into Linux from z/VM & z/OS the > > 'tcpdump -I hsi0 ip proto \\icmp' on the Linux only shows the request > > received. No reply is sent from Linux back. The z/OS, z/VM, and Linux > have > > OSAs to reach the world. The OSA address for z/VM & z/OS is what comes > > across the hsi0 ping. > > > > We need to bring customer requests in to the OSA - VSWITCH - Linux - hsi0 > - > > z/OS and back through the process. We are quite perplexed. Any help > would > > be greatly appreciated. Just looking at this again. Since you didn't see the ICMP ECHO RESPONSE go back out on the hsi0 interface, you can come to two conclusion 1. The ECHO RESPONSE went out over another interface, OR 2. The ECHO RESPONSE didn't go out at all. Remove the -I hsi0 filter and see what you see. It is also interesting to see the ORIGIN IP address in incoming the ECHO REQUEST. This may have something to do with the IP addresses z/VM and z/OS use on their outbound packets. Alan Altmark z/VM and Linux on System z Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 alan_altm...@us.ibm.com IBM Endicott -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets Not working z/Linux to z/VM & z/OS
10.90.1.16 UGHO 00 OSD2 10.90.23.52/32 10.90.1.15 UGHO 00 OSD2 10.90.23.53/32 10.90.1.16 UGHO 00 OSD0 10.90.23.53/32 10.90.1.15 UGHO 00 OSD0 10.90.23.53/32 10.90.1.16 UGHO 00 OSD2 10.90.23.53/32 10.90.1.15 UGHO 00 OSD2 10.90.23.54/32 10.90.1.16 UGHO 00 OSD0 10.90.23.54/32 10.90.1.15 UGHO 00 OSD0 10.90.23.54/32 10.90.1.16 UGHO 00 OSD2 10.90.23.54/32 10.90.1.15 UGHO 00 OSD2 10.90.23.55/32 10.90.1.16 UGHO 00 OSD0 10.90.23.55/32 10.90.1.15 UGHO 00 OSD0 10.90.23.55/32 10.90.1.16 UGHO 00 OSD2 10.90.23.55/32 10.90.1.15 UGHO 00 OSD2 10.90.23.56/32 10.90.1.16 UGHO 00 OSD0 10.90.23.56/32 10.90.1.15 UGHO 00 OSD0 10.90.23.56/32 10.90.1.16 UGHO 00 OSD2 10.90.23.56/32 10.90.1.15 UGHO 00 OSD2 10.90.23.57/32 10.90.1.16 UGHO 00 OSD0 10.90.23.57/32 10.90.1.15 UGHO 00 OSD0 10.90.23.57/32 10.90.1.16 UGHO 00 OSD2 10.90.23.57/32 10.90.1.15 UGHO 00 OSD2 10.90.23.58/32 10.90.1.16 UGHO 01 OSD0 10.90.23.58/32 10.90.1.15 UGHO 00 OSD0 10.90.23.58/32 10.90.1.16 UGHO 00 OSD2 10.90.23.58/32 10.90.1.15 UGHO 00 OSD2 10.90.24.0/24 10.90.1.18 UGO 00 OSD0 10.90.24.0/24 10.90.1.17 UGO 00 OSD0 10.90.24.0/24 10.90.1.18 UGO 00 OSD2 10.90.24.0/24 10.90.1.17 UGO 00 OSD2 10.90.24.1/32 10.90.1.18 UGHO 03 OSD0 10.90.24.1/32 10.90.1.17 UGHO 00 OSD0 10.90.24.1/32 10.90.1.18 UGHO 00 OSD2 10.90.24.1/32 10.90.1.17 UGHO 00 OSD2 10.90.24.10/32 10.90.1.18 UGHO 00 OSD0 10.90.24.10/32 10.90.1.17 UGHO 00 OSD0 10.90.24.10/32 10.90.1.18 UGHO 00 OSD2 10.90.24.10/32 10.90.1.17 UGHO 00 OSD2 10.90.25.0/24 10.90.1.13 UGO 00 OSD0 10.90.25.0/24 10.90.1.14 UGO 00 OSD0 10.90.25.0/24 10.90.1.13 UGO 00 OSD2 10.90.25.0/24 10.90.1.14 UGO 00 OSD2 10.90.25.1/32 10.90.1.13 UGHO 00 OSD0 10.90.25.1/32 10.90.1.14 UGHO 00 OSD0 10.90.25.1/32 10.90.1.13 UGHO 00 OSD2 10.90.25.1/32 10.90.1.14 UGHO 00 OSD2 10.233.254.0/2410.90.1.12 UGO 00 OSD0 10.233.254.0/2410.90.1.11 UGO 00 OSD0 10.233.254.0/2410.90.1.12 UGO 00 OSD2 10.233.254.0/2410.90.1.11 UGO 00 OSD2 10.233.254.251/32 10.90.1.12 UGHO 01 OSD0 10.233.254.251/32 10.90.1.11 UGHO 00 OSD0 10.233.254.251/32 10.90.1.12 UGHO 00 OSD2 10.233.254.251/32 10.90.1.11 UGHO 00 OSD2 127.0.0.1/32 0.0.0.0 UH 50 LOOPBACK 172.18.18.0/24 10.90.1.1 UGO 00 OSD0 172.18.18.0/24 10.90.1.1 UGO 00 OSD2 Thanks for any suggestions/help. Kyle -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Mark Post Sent: Sunday, January 30, 2011 7:52 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets Not working z/Linux to z/VM & z/OS >>> On 1/30/2011 at 03:19 AM, Naspa wrote: > Greetings All! > > We have been struggling trying to get our z/VM & z/OS systems to talk via > hipersockets to our RHEL5.3 Linux (under z/VM 6.1). We can ping out from > Linux to VM or to z/OS, but when we ping into Linux from z/VM & z/OS the > 'tcpdump -I hsi0 ip proto \\icmp' on the Linux only shows the request > received. No reply is sent from Linux back. The z/OS, z/VM, and Linux have > OSAs to reach the world. The OSA address for z/VM & z/OS is what comes > across the hsi0 ping. > > We need to bring customer requests in to the OSA - VSWITCH - Linux - hsi0 - > z/OS and back through the process. We are quite perplexed. Any help would > be greatly appreciated. What IP addresses and subnets are defined for each interface? What are they for z/VM and z/OS? Are the hsi0 interfaces real HiperSockets, or virtual ones? What does your routing table look like on all systems? Mark Post -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Hipersockets Not working z/Linux to z/VM & z/OS
>>> On 1/30/2011 at 03:19 AM, Naspa wrote: > Greetings All! > > We have been struggling trying to get our z/VM & z/OS systems to talk via > hipersockets to our RHEL5.3 Linux (under z/VM 6.1). We can ping out from > Linux to VM or to z/OS, but when we ping into Linux from z/VM & z/OS the > 'tcpdump -I hsi0 ip proto \\icmp' on the Linux only shows the request > received. No reply is sent from Linux back. The z/OS, z/VM, and Linux have > OSAs to reach the world. The OSA address for z/VM & z/OS is what comes > across the hsi0 ping. > > We need to bring customer requests in to the OSA - VSWITCH - Linux - hsi0 - > z/OS and back through the process. We are quite perplexed. Any help would > be greatly appreciated. What IP addresses and subnets are defined for each interface? What are they for z/VM and z/OS? Are the hsi0 interfaces real HiperSockets, or virtual ones? What does your routing table look like on all systems? Mark Post -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets Not working z/Linux to z/VM & z/OS
I forgot to mention that we can ping via hsi0 to /from either of the Linux guests. -Original Message- From: Naspa Sent: Sunday, January 30, 2011 1:19 AM To: 'Linux on 390 Port' Subject: Hipersockets Not working z/Linux to z/VM & z/OS Greetings All! We have been struggling trying to get our z/VM & z/OS systems to talk via hipersockets to our RHEL5.3 Linux (under z/VM 6.1). We can ping out from Linux to VM or to z/OS, but when we ping into Linux from z/VM & z/OS the 'tcpdump -I hsi0 ip proto \\icmp' on the Linux only shows the request received. No reply is sent from Linux back. The z/OS, z/VM, and Linux have OSAs to reach the world. The OSA address for z/VM & z/OS is what comes across the hsi0 ping. We need to bring customer requests in to the OSA - VSWITCH - Linux - hsi0 - z/OS and back through the process. We are quite perplexed. Any help would be greatly appreciated. Thank you, Kyle Stewart -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: Hipersockets on Redhat
On Thu, 2009-04-09 at 13:18 -0400, Bauer, Bobby (NIH/CIT) [E] wrote: > echo 0.0.710c,0.0.710d,0.0.710e > group On Thu, 2009-04-09 at 10:34 -0400, Bauer, Bobby (NIH/CIT) [E] wrote: > SUBCHANNELS=0.0.710C,0.0.710D,0.0.710E You should change SUBCHANNELS value to lowercase: SUBCHANNELS=0.0.710c,0.0.710d,0.0.710e This should fix the Problem. Regards, Oliver Paukstadt -- Oliver Paukstadt -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Hipersockets on Redhat
On Thu, 9 Apr 2009 13:18:03 -0400 "Bauer, Bobby (NIH/CIT) [E]" wrote: > Seems Redhat doesn't have 'hwup' or /etc/sysconfig/hardware however the > 'lsqeth' did not know about the hipersocket addresses. Red Hat equivalent is ifup and I would guess that /etc/sysconfig-network-scripts/ifcfg-foo is the other needed bit. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Hipersockets on Redhat
Seems Redhat doesn't have 'hwup' or /etc/sysconfig/hardware however the 'lsqeth' did not know about the hipersocket addresses. So I added them dynamically: cd /sys/bus/ccwgroup/drivers/qeth echo 0.0.710c,0.0.710d,0.0.710e > group cd 0.0.710c echo 1 > online ifup hsi0 and it worked. I noticed the last character in the echo command above was in lower case and in ifcfg-hsi0 I had upper case. Change them to lower case, rebooted and they came online as expected. Bobby Bauer Center for Information Technology National Institutes of Health Bethesda, MD 20892-5628 301-594-7474 -Original Message- From: Linux on 390 Port [mailto:linux-...@vm.marist.edu] On Behalf Of Rob van der Heij Sent: Thursday, April 09, 2009 11:00 AM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets on Redhat Also try whether "hwup" of the device helps to get it online. On Thu, Apr 9, 2009 at 4:39 PM, Harder, Pieter wrote: > MODULE='qeth' That one finally turned out for me to be missing... after I have tried dozens of bizarre things with Mark Post at the strangest moments in our respective nights (I see how this phrase could give the wrong impression) Rob -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Hipersockets on Redhat
Also try whether "hwup" of the device helps to get it online. On Thu, Apr 9, 2009 at 4:39 PM, Harder, Pieter wrote: > MODULE='qeth' That one finally turned out for me to be missing... after I have tried dozens of bizarre things with Mark Post at the strangest moments in our respective nights (I see how this phrase could give the wrong impression) Rob -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Hipersockets on Redhat
What does lsqeth tell you? What do you have in /etc/sysconfig/hardware? Mine is hwcfg-hsi-bus-ccw-0.0 CCW_CHAN_IDS='0.0.0c03 0.0.0c04 0.0.0c05' CCW_CHAN_MODE='' CCW_CHAN_NUM='3' LCS_LANCMD_TIMEOUT='' MODULE='qeth' MODULE_OPTIONS='' QETH_IPA_TAKEOVER='0' QETH_LAYER2_SUPPORT='0' QETH_OPTIONS='' SCRIPTDOWN='hwdown-ccw' SCRIPTUP='hwup-ccw' SCRIPTUP_ccw='hwup-ccw' SCRIPTUP_ccwgroup='hwup-qeth' STARTMODE='auto' Best regards, Pieter Harder pieter.har...@brabantwater.nl tel +31-73-6837133 / +31-6-47272537 -Oorspronkelijk bericht- Van: Linux on 390 Port [mailto:linux-...@vm.marist.edu] Namens Bauer, Bobby (NIH/CIT) [E] Verzonden: donderdag 9 april 2009 16:34 Aan: LINUX-390@VM.MARIST.EDU Onderwerp: Hipersockets on Redhat I think I need somebody to look over my shoulder. I just built a new server and cannot get the hipersockets to come online. They are working just fine on the one one other server I have them setup on using addresses 7104-7106. They are define in HCD as: 7100,48 IQD In user direct: DEDICATE 710C 710C DEDICATE 710D 710D DEDICATE 710E 710E In /etc/modprobe.conf: alias hsi0 qeth options dasd_mod dasd=201-219 alias scsi_hostadapter zfcp in /etc/sysconfig/network-scripts/ifcfg-hsi0: DEVICE=hsi0 BOOTPROTO=static ONBOOT=yes NETTYPE=qeth MTU=8192 SUBCHANNELS=0.0.710C,0.0.710D,0.0.710E IPADDR=10.1.160.43 NETMASK=255.255.255.255 OPTIONS='fake_ll=1' A 'vmcp q osa' displays: OSA 710C ON OSA 710C SUBCHANNEL = 0015 710C DEVTYPE HIPER CHPID 20 IQD 710C QDIO-ELIGIBLE QIOASSIST-ELIGIBLE OSA 710D ON OSA 710D SUBCHANNEL = 0016 710D DEVTYPE HIPER CHPID 20 IQD 710D QDIO-ELIGIBLE QIOASSIST-ELIGIBLE OSA 710E ON OSA 710E SUBCHANNEL = 0017 710E DEVTYPE HIPER CHPID 20 IQD 710E QDIO-ELIGIBLE QIOASSIST-ELIGIBLE But ifup hsi0 tells me: qeth device hsi0 does not seem to be present, delaying initialization. All this looks like the server that is working. Anybody see anything I've missed? Thanks Bobby Bauer Center for Information Technology National Institutes of Health Bethesda, MD 20892-5628 301-594-7474 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 Brabant Water N.V. Postbus 1068 5200 BC 's-Hertogenbosch http://www.brabantwater.nl Handelsregister: 16005077 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Hipersockets
Mark Post wrote: On Thu, Jul 26, 2007 at 3:56 PM, in message <[EMAIL PROTECTED]>, "Ayer, Paul W" <[EMAIL PROTECTED]> wrote: -snip- Should I add a DNS z/Linux system inside the z/VM as a first check that would resolve the same name to a 12 Then if not found there somehow go out to the network DNS. Should I try to use the old HOSTFILEs type datasets? How have others addresses this? When I was supporting midrange Linux systems, we would always assign different DNS names to different interfaces on the same system. Always. You don't want to mess around with adding entries to /etc/hosts if you can avoid it. I always have DNS (ISC BIND) set up. I never use /etc/hosts. OTOH I only do toys. I'm with Mark on different names for different interfaces. I don't know a reason to have my.example.com on two networks. Two hosts on the same network, maybe, but not on different networks. -- Cheers John -- spambait [EMAIL PROTECTED] [EMAIL PROTECTED] Please do not reply off-list -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Hipersockets
> Doesn't this second entry conflict with the other CNAME entry? Does this > mean you have to have two DNS servers to implement it this way? If it's in a different zone file, no, because if a user specifies an unqualified vm1foo01 (which I think is what he's asking for), the current zone context is assumed, and all proceeds as planned. This approach (coupled with our convention of hosts that need to do things like this having a caching DNS installed on the host) also allows strategic poisoning of the cache to do this sort of magic w/o using host files. Also, it depends on what DNS implementation you use. Many take the last entry found, some take the first. If you're using bind, I think (no manuals avail at the moment) it takes the first one. If you're really clever, you can also use split horizon support to do something similar (ie, give a completely different answer depending on who asks), but not all DNS implementations do a good job of supporting that. > It seems like it would be ok to direct traffic to > vm1foo01-his0.guest.comeven when the host thinks of itself as > vm1foo01, but other than confusion > are there any dragons there? Depends on whether vm1foo01-hsi0 is actually accessible from the host in question. This is policy-based routing at it's finest: technically it may work, but not politically. It's a lot harder if the host has to handle mail. The MTA configs in modern mail delivery will get very cranky and complicated if the hostname doesn't match what is expected. But Mark's other comment is right on -- long term, name the individual interfaces, and get the users to use the name for the interface they want if they want policy-based path selection in IP networks. The above scheme gets complex if there are more than a few aliases. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Hipersockets
>>> On Fri, Jul 27, 2007 at 3:04 PM, in message <[EMAIL PROTECTED]>, Robert J Brenneman <[EMAIL PROTECTED]> wrote: -snip- > Doesn't this second entry conflict with the other CNAME entry? Does this > mean you have to have two DNS servers to implement it this way? In the general case you only do this if you want different systems to interpret the CNAME differently, so yes. > It seems like it would be ok to direct traffic to > vm1foo01-his0.guest.com even when the host thinks of itself as > vm1foo01, but other than confusion > are there any dragons there? Not really. What we used it for mainly was to separate the system management interface from the backup interface, from the interface(s) serving the business purpose of the server. So, we would do "ssh hostname-m" "ssh hostname-drac," etc. You're not really referring to a system by a "different hostname" because the only thing that is going on is (DNS) name resolution, pointing you to the proper NIC. Mark Post -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: hipersockets + route update command allows one DNS name
> After I sent the note I was thinking that too. Now how to I say FTP XYZ > and make it use one path or the other. The method I suggested allows that exact thing to work seamlessly. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Hipersockets
On 7/27/07, David Boyes <[EMAIL PROTECTED]> wrote: > > > That's what CNAME records in the DNS are for. > > vm1foo01-eth0 IN A 123.45.67.89 > vm1foo01-hsi0 IN A 10.11.12.13 > vm1foo01IN CNAME vm1foo01-eth0 > > In the reverse zone files, we create: > > 123.45.67.zone: > 89 IN PTR vm1foo01-eth0.guest.com. > > 10.11.12.zone: > 13 IN PTR vm1foo01-his0.guest.com. Another host might create a DNS entry for vm1foo01, but want to refer to > the hipersocket interface only. They create: > > vm1foo01IN CNAME vm1foo01-hsi0.guest.com > > and any reference they make to vm1foo01 goes to the hipersocket > interface. > > Doesn't this second entry conflict with the other CNAME entry? Does this mean you have to have two DNS servers to implement it this way? It seems like it would be ok to direct traffic to vm1foo01-his0.guest.comeven when the host thinks of itself as vm1foo01, but other than confusion are there any dragons there? -- Jay Brenneman -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: hipersockets + route update command allows one DNS name
Thanks Mark, After I sent the note I was thinking that too. Now how to I say FTP XYZ and make it use one path or the other. Oh well, the users and applications will need to be really flexible since we plan to be able to move an instance between more than one CPU so sometimes the Hipersockets DNS name will work and sometimes they will need to use the wide area DNS name instead. Thanks to everyone all for your input. Paul -Original Message- From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On Behalf Of Mark Post Sent: Friday, July 27, 2007 2:11 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: hipersockets + route update command allows one DNS name >>> On Fri, Jul 27, 2007 at 11:08 AM, in message <[EMAIL PROTECTED]>, "Ayer, Paul W" <[EMAIL PROTECTED]> wrote: > This seems to work fine. > Please tell me where I may be in error or have a possible problem > waiting to happen. -snip- > pings and trace routes now work fine via the path they should for the > same DNS name ... > ping -c 1 -I eth0 agzls013 > PING agzls013 (192.168.127.5) from 192.168.32.139 eth0: 56(84) bytes of > data. > 64 bytes from agzls013 (192.168.127.5): icmp_seq=0 ttl=59 time=2.44 ms > --- agzls013 ping statistics --- > 1 packets transmitted, 1 received, 0% packet loss, time 0ms > rtt min/avg/max/mdev = 2.446/2.446/2.446/0.000 ms, pipe 2 > ping -c 1 -I hsi0 agzls013 > PING agzls013 (192.168.127.5) from 12.1.102.9 hsi0: 56(84) bytes of > data. > 64 bytes from agzls013 (192.168.127.5): icmp_seq=0 ttl=64 time=1.07 ms > --- agzls013 ping statistics --- > 1 packets transmitted, 1 received, 0% packet loss, time 0ms > rtt min/avg/max/mdev = 1.078/1.078/1.078/0.000 ms, pipe 2 Things are working for your artificial test because you're forcing the ping command to use a particular interface. What will force the application to do that? (Answer: nothing.) Give the HiperSocket interfaces a different name and use those. That really is how things are supposed to work. Mark Post -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: hipersockets + route update command allows one DNS name
>>> On Fri, Jul 27, 2007 at 11:08 AM, in message <[EMAIL PROTECTED]>, "Ayer, Paul W" <[EMAIL PROTECTED]> wrote: > This seems to work fine. > Please tell me where I may be in error or have a possible problem > waiting to happen. -snip- > pings and trace routes now work fine via the path they should for the > same DNS name ... > ping -c 1 -I eth0 agzls013 > PING agzls013 (192.168.127.5) from 192.168.32.139 eth0: 56(84) bytes of > data. > 64 bytes from agzls013 (192.168.127.5): icmp_seq=0 ttl=59 time=2.44 ms > --- agzls013 ping statistics --- > 1 packets transmitted, 1 received, 0% packet loss, time 0ms > rtt min/avg/max/mdev = 2.446/2.446/2.446/0.000 ms, pipe 2 > ping -c 1 -I hsi0 agzls013 > PING agzls013 (192.168.127.5) from 12.1.102.9 hsi0: 56(84) bytes of > data. > 64 bytes from agzls013 (192.168.127.5): icmp_seq=0 ttl=64 time=1.07 ms > --- agzls013 ping statistics --- > 1 packets transmitted, 1 received, 0% packet loss, time 0ms > rtt min/avg/max/mdev = 1.078/1.078/1.078/0.000 ms, pipe 2 Things are working for your artificial test because you're forcing the ping command to use a particular interface. What will force the application to do that? (Answer: nothing.) Give the HiperSocket interfaces a different name and use those. That really is how things are supposed to work. Mark Post -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Hipersockets
> Since this is an internal only network the ip address should not matter > at all (my thought anyway) Oh, it matters a lot. That's why there are reserved address ranges of differing sizes for this kind of stuff. You *will* have problems if you do this. Classic example: Sun used to publish examples that used actual addresses from their network in their manuals. A lot of people set up their networks using the example addresses -- and then connected to the internet, where a lot of people suddenly had to do mass renumbering because their machines were attempting to boot from machines at Sun...8-) > but I guess if were to reserve a > 10.*.*.* (* something) for z/linux only then that would be ok. Use a /24 from 192.168 or 172.28 instead. Wastes less space, and a lot of large corps use net 10 for internal space. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Hipersockets
> Lots of systems get really upset if you try to assign multiple hostnames > to > the same tcpip stack. That's what CNAME records in the DNS are for. I create a DNS forward and reverse for each interface (canonical name-interface, eg foo-eth0, foo-hsi0, etc.). The "hostname" that I configure as /etc/hostname is a CNAME pointing to the appropriate interface, which makes it nicely portable. Most of the complaining about multiple entries pointing to an address is caused by the forward entry not matching the reverse entry. CNAMEs force another DNS recursion to get the info of the actual interface, which has forward and reverse matching, and everyone's happy. If it's mail, you need MX and SPF records as well. If I want a pointer to a host from the same zone or another zone, I can use a CNAME to the FQDN of the interface I want to use. It's a little more work, but it's really flexible, and helps a lot when you start doing VIPA and other tricks. Example: Linux guest vm1foo01.guest.com has two interfaces, one external Ethernet, one internal hipersocket. In my guest.com DNS zone I have: vm1foo01-eth0 IN A 123.45.67.89 vm1foo01-hsi0 IN A 10.11.12.13 vm1foo01IN CNAME vm1foo01-eth0 In the reverse zone files, we create: 123.45.67.zone: 89 IN PTR vm1foo01-eth0.guest.com. 10.11.12.zone: 13 IN PTR vm1foo01-his0.guest.com. Users in guest.com refer to vmfoo01, and they get the vm1foo01-eth0 address. Any host that actually checks forward and reverse matches has to follow the CNAME pointer to the canonical entry, and those entries always match correctly. Another host might create a DNS entry for vm1foo01, but want to refer to the hipersocket interface only. They create: vm1foo01IN CNAME vm1foo01-hsi0.guest.com and any reference they make to vm1foo01 goes to the hipersocket interface. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Hipersockets
Lots of systems get really upset if you try to assign multiple hostnames to the same tcpip stack. I set up two different DNS domains for this type of configuration. Each host on the network has only one hostname: Linux1 Linux2 etc Each host is actually a member of two domains: my.external.net: 192.168.70.0/24 my.hipersocket.net: 10.1.1.0/24 So, Linux1 has only one hostname - Linux1 - but it has two fully qualified hostnames: Linux1.my.external.net Linux1.my.hipersocket.net Both domains (zones?) are defined in the DNS that is out on the 192.168.70network, so you alter the domain part of the FQDN to control which network you use to get to an adjacent system. It seems to work well. -- Jay Brenneman -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Hipersockets
Hi Mike, Since this is an internal only network the ip address should not matter at all (my thought anyway), plus if we were to use a real routable address that could become a problem if a session to that address were made over the OSA (not hipersockets path) ... ??? but I guess if were to reserve a 10.*.*.* (* something) for z/linux only then that would be ok. -Original Message- From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On Behalf Of Michael MacIsaac Sent: Friday, July 27, 2007 8:50 AM To: LINUX-390@VM.MARIST.EDU Subject: Re: Hipersockets Paul, > setup the hipersocket network as 12. for testing and so it stands out Would it be better to use 10.? 12. is a real class A IP address range. 10. is the private class A range (RFC 1918 I believe). "Mike MacIsaac" <[EMAIL PROTECTED]> (845) 433-7061 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Hipersockets
Paul, > setup the hipersocket network as 12. for testing and so it stands out Would it be better to use 10.? 12. is a real class A IP address range. 10. is the private class A range (RFC 1918 I believe). "Mike MacIsaac" <[EMAIL PROTECTED]> (845) 433-7061 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Hipersockets
>>> On Thu, Jul 26, 2007 at 3:56 PM, in message <[EMAIL PROTECTED]>, "Ayer, Paul W" <[EMAIL PROTECTED]> wrote: -snip- > Should I add a DNS z/Linux system inside the z/VM as a first check that > would resolve the same name to a 12 > Then if not found there somehow go out to the network DNS. Should I try > to use the old HOSTFILEs type datasets? > > How have others addresses this? When I was supporting midrange Linux systems, we would always assign different DNS names to different interfaces on the same system. Always. You don't want to mess around with adding entries to /etc/hosts if you can avoid it. Mark Post -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Hipersockets Conundrum Revisited
On Monday, 07/09/2007 at 11:10 AST, Kim Goldenberg <[EMAIL PROTECTED]> wrote: > HOME > xxx.yyy.1.101 HIPER50 Make sure you are using chpid 50 on Linux. 1. CP Q V xxx (where xxx is the vdev of the HiperSocket). Note the SUBCHANNEL number. 2. CP D SCHIB nnn (where nnn is the subchannel number) 3. Look at the chpid number. Does it match? > (There is nothing explicit in routing information, we are using OMPROUTE > on z/OS) You need to display the routing table on z/OS. Since you're not [shouldn't be] running OSPF or RIP on the HiperSocket there will be an INTERFACE statement in omproute's configuration. If there isn't, then it is defaulting and omproute defaults are to be avoided at all costs. Alan Altmark z/VM Development IBM Endicott -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Hipersockets Conundrum Revisited
Alan Altmark wrote: On Monday, 07/09/2007 at 11:10 AST, Kim Goldenberg You need to display the routing table on z/OS. Since you're not [shouldn't be] running OSPF or RIP on the HiperSocket there will be an INTERFACE statement in omproute's configuration. If there isn't, then it is defaulting and omproute defaults are to be avoided at all costs. That was the problem. Just before I read this I, an explicit route was put in, and all works as expected now! Thanks for the hints. Kim -- Kim Goldenberg Systems Programmer I State of NJ - OIT 609-777-3722 [EMAIL PROTECTED] [EMAIL PROTECTED] -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
Re: Hipersockets Conundrum Revisited
> > What are the IP addresses and subnet masks on the z/OS HiperSocket > interfaces? > ; *** > ; HIPERSOCKET CHIPID 50 > ; *** > DEVICE IUTIQD50 MPCIPA > LINK HIPER50 IPAQIDIO IUTIQD50 > HOME >zzz.yyy.72.82 OSA2 >xxx.yyy.11.1VLINK1 >xxx.yyy.1.22OSA1 >xxx.yyy.1.13OSA3 >xxx.yyy.1.101 HIPER50 >PRIMARYINTERFACE VLINK1 > (There is nothing explicit in routing information, we are using OMPROUTE > on z/OS) Hmm. This may be the problem. Since your HS link is in the same subnet as the OSA1 and OSA3 links, the routing information is going to be important; you need an explicit route on the z/OS side that points at the hipersocket interface for the portion of the subnet that contains the Linux guests. Can you dump the current routing table on z/OS? Also, the subnet mask is going to be important. What mask are you using for the HIPER50 interface? If it's not a routing problem, I'd have to agree with Alan -- they're not on the same HS chpid. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
