Re: z16 CPU type/model number
Hi Mark, that would be 3931. so long, Carsten From: Linux on 390 Port on behalf of Mark Post Sent: Wednesday, April 13, 2022 8:45 PM To: LINUX-390@VM.MARIST.EDU Subject: [EXTERNAL] z16 CPU type/model number All, Now that the z16 has been announced, I'd like to add the machine type to the cputype script. Does anyone know what that model number is? Thanks, 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://www2.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://www2.marist.edu/htbin/wlvindex?LINUX-390
Re: zKVM on zpdt
Hi Tito, I am sorry that this time I don't have a workaround. The installer code assumes diag 308 IPL functions to be available, however it is'nt on zPDT. I believe this should be fixed in the installer. The Linux kernel propagates the error to the script, and it fails to handle it. so long, Carsten From: Tito Garrido To: LINUX-390@VM.MARIST.EDU Date: 05/10/2015 23:57 Subject:Re: zKVM on zpdt Sent by:Linux on 390 Port Hi Carsten, Thanks for your inputs. I have formatted from another linux and it worked. Now I got another error on chreipl step: *2015-10-02 17:16:24,106 - model.installfunctions - DEBUG - Zipl output is: Using config file '/etc/zipl.conf'Building bootmap in '/boot'Building menu 'zipl-automatic-menu'Adding #1: IPL section 'linux' (default)Preparing boot device: dasda (5001).Done.2015-10-02 17:16:24,111 - model.installfunctions - DEBUG - Zipl error output is:2015-10-02 17:16:24,130 - model.installfunctions - DEBUG - Zipl bootName is: dasda2015-10-02 17:16:24,187 - model.installfunctions - DEBUG - chreipl output is:2015-10-02 17:16:24,191 - model.installfunctions - DEBUG - chreipl error output is: chreipl: Could not open "reipl/ccw/loadparm" (Permission denied)2015-10-02 17:16:24,194 - model.installfunctions - ERROR - Error running chreipl2015-10-02 17:16:24,198 - model.installfunctions - CRITICAL - Failed installSystem2015-10-02 17:16:24,201 - model.installfunctions - CRITICAL - EXCEPTION:2015-10-02 17:16:24,204 - model.installfunctions - CRITICAL - Error running chreipl2015-10-02 17:16:24,223 - model.installfunctions - CRITICAL - Stacktrace:Traceback (most recent call last): File "/opt/ibm/kvmibm-installer/model/installfunctions.py", line 377, in installSystem installBootloader(diskSelected, rootDir, bootDev, rootDev, swapDev) File "/opt/ibm/kvmibm-installer/model/installfunctions.py", line 1101, in installBootloader raise RuntimeError('Error running chreipl')RuntimeError: Error running chreipl2015-10-02 17:16:24,625 - controller.controller - CRITICAL - ZKVMError: [['KVMIBMIN70500', 'Error while installing packages.'], ('INSTALLER', 'INSTALLSYSTEM', 'INSTALL_MSG')](END)* Probably is another issue with zPDT. On Mon, Oct 5, 2015 at 5:30 AM, Carsten Otte wrote: > > Tito, > > I think you ran into a Linux dasd device driver problem. The issue is, that > the awsckd emulator does not support (and does not advertiese) the prefix > command. This is perfectly fine for a 3390 device, however Linux does issue > this command on 3390. The issue has been reported and is fixed in upstream > Linux. I think the bugfix needs to be backported to your level of code. > As a workaround, you can bring up another linux that is either old enough > for not having the bug or new enough to have the fix, and do dasdfmt on > your dasd. Using the dasd (partitioning, creating and mounting filesystems) > works flawless from your kernel as far as I can tell. > > so long, > Carsten > -- > Carsten Otte > IBM Deutschland R&D > Firmware Development > > > > From: Tito Garrido > To: LINUX-390@VM.MARIST.EDU > Date: 02/10/2015 18:30 > Subject:Re: zKVM on zpdt > Sent by:Linux on 390 Port > > > > FYI: To make it work I had to format the dasds on an older Linux like > RHEL6. > > On Fri, Oct 2, 2015 at 10:47 AM, Tito Garrido > wrote: > > > Hi Folks, > > > > I have already asked this question on z1090 mail list but it may be > > interesting for other people here that would like to try zKVM and also > you > > may know the answer :) > > > > > > I am trying to install zKVM on zpdt but it is not able to run dasdfmt > > during the installation: > > > > > > 2015-10-02 02:13:15,704 - program - INFO - Running... /sbin/dasdfmt -y -d > > cdl -b 4096 /dev/dasda > > 2015-10-02 02:13:15,734 - program - INFO - /sbin/dasdfmt: (invalidate > > first track) IOCTL BIODASDFMT failed. (Input/output error) > > > > Any clue? > > > > I have already tried to run CPFMTXA on a z/VM instance and run the > > installation again but no success... > > > > > > > > Regards, > > Tito > > > > -- > > > > Linux User #387870 > > . > > _/_õ|__| > > ..º[ .-.___.-._| . . . . > > .__( o)__( o).:___ > > > > > > -- > > Linux User #387870 > . > _/_õ|__| > ..º[ .-.___.-._| . . . . > .__( o)__( o).:___ > > -- > For LINUX-390 subscribe / signoff / archive access instructions, > send ema
Re: zKVM on zpdt
Tito, I think you ran into a Linux dasd device driver problem. The issue is, that the awsckd emulator does not support (and does not advertiese) the prefix command. This is perfectly fine for a 3390 device, however Linux does issue this command on 3390. The issue has been reported and is fixed in upstream Linux. I think the bugfix needs to be backported to your level of code. As a workaround, you can bring up another linux that is either old enough for not having the bug or new enough to have the fix, and do dasdfmt on your dasd. Using the dasd (partitioning, creating and mounting filesystems) works flawless from your kernel as far as I can tell. so long, Carsten -- Carsten Otte IBM Deutschland R&D Firmware Development From: Tito Garrido To: LINUX-390@VM.MARIST.EDU Date: 02/10/2015 18:30 Subject:Re: zKVM on zpdt Sent by:Linux on 390 Port FYI: To make it work I had to format the dasds on an older Linux like RHEL6. On Fri, Oct 2, 2015 at 10:47 AM, Tito Garrido wrote: > Hi Folks, > > I have already asked this question on z1090 mail list but it may be > interesting for other people here that would like to try zKVM and also you > may know the answer :) > > > I am trying to install zKVM on zpdt but it is not able to run dasdfmt > during the installation: > > > 2015-10-02 02:13:15,704 - program - INFO - Running... /sbin/dasdfmt -y -d > cdl -b 4096 /dev/dasda > 2015-10-02 02:13:15,734 - program - INFO - /sbin/dasdfmt: (invalidate > first track) IOCTL BIODASDFMT failed. (Input/output error) > > Any clue? > > I have already tried to run CPFMTXA on a z/VM instance and run the > installation again but no success... > > > > Regards, > Tito > > -- > > Linux User #387870 > . > _/_õ|__| > ..º[ .-.___.-._| . . . . > .__( o)__( o).:___ > -- Linux User #387870 . _/_õ|__| ..º[ .-.___.-._| . . . . .__( o)__( o).:___ -- 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: Linux file updates by timestamp and userid
How about this open source tool? http://stackoverflow.com/questions/10527936/using-inotify-to-keep-track-of-all-files-in-a-system with kind regards Carsten Otte System z firmware development / Boeblingen lab --- Every revolution was first a thought in one man's mind; and when the same thought occurs to another man, it is the key to that era. - Ralph Waldo Emerson, Essays: First Series, 1841 "Shan, Rita" To Sent by: Linux on LINUX-390@vm.marist.edu, 390 Port cc Subject Linux file updates by timestamp and userid 13.03.2014 23:32 Please respond to Linux on 390 Port Could anyone kindly provide information on how we can monitor/log zLinux file updates by timestamp and by user ID? We have a number of staff maintaining zLinux system all with sudo privilege, we need to have a way to track file updates by date/time/user-ID. Does AIDE provides these kind of detailed level information? What kind of overhead it will generate if we turned it on? Is there an inexpensive vendor tool for this? Any help is greatly appreciated Rita Email transmitted across the Internet is normally not protected and may be intercepted and viewed by others. Therefore, you should refrain from sending any confidential or private information via unsecured email to PenFed. We will not ask you to send confidential information to us via email, such as your logon ID, password, account numbers, or Social Security number. We prohibit our employees from sending confidential information to you via email that is not encrypted. The recommended document submission method is FAX; a partial list of generic fax numbers can be found < https://www.penfed.org/aboutUs/contactUs.asp#fax> here< https://www.penfed.org/aboutUs/contactUs.asp#fax>.< https://www.penfed.org/aboutUs/contactUs.asp#fax> -- 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: Layer 2 frames passing through a Linux bridge get dropped before leaving the mainframe box
This setup won't work, because Linux negotiates its mac address with the OSA, and cannot send frames from another mac. You could use ip forwarding, and have Linux route on layer 3. This should work, as long as you use the OSA in layer 2 mode. with kind regards Carsten Otte System z firmware development / Boeblingen lab --- Every revolution was first a thought in one man's mind; and when the same thought occurs to another man, it is the key to that era. - Ralph Waldo Emerson, Essays: First Series, 1841 "Pavelka, Tomas" To Sent by: Linux on LINUX-390@vm.marist.edu, 390 Port cc Subject Layer 2 frames passing through a Linux bridge get dropped before 20.02.2014 09:04 leaving the mainframe box Please respond to Linux on 390 Port We have a problem where frames that pass through a Linux bridge do not reach the gateway outside of the mainframe box. We have set up an experiment that reproduces the problem, which looks like this: (LINUX1) - - (LINUXBR) - - OSA - gateway The problem is that in this setup we cannot ping the gateway. But, under a different setup: (LINUX1) - - (LINUXBR) - - (LINUX2) Both LINUX1 and LINUX2 can communicate. Moreover, LINUX2 can ping the gateway (the OSA card is still connected to the public vswitch, I just did not put it in the picture). Some more details that may be important: - Both public and private vswitch are layer 2 - LINUXBR runs RHEL 6 and uses bridge-utils to create the bridge - private vswitch is not connected to any OSA card We have played with TCPDUMP and found that ARP (broadcast) packets do reach the gateway and come back, but ping's ICMP (unicast) packets get dropped. This led us to the following hypothesis: If there is a unicast packet originating from a MAC address not known to public vswitch, it gets dropped somewhere on the way between LINUXBR and the gateway. Does anyone know any settings that could affect filtering done either by the vswitch or by the OSA card? We asked our hardware people but they did not know of anything that could cause the problems. But a more targeted question could help if we knew what to ask for. Any debugging tips will be much appreciated. Thanks, Tomas Tomas Pavelka CA Technologies Sr Software Engineer Tel: +420226207796 tomas.pave...@ca.com <mailto:tomas.pave...@ca.com>[cid:image001.gif@01CF2E1A.CF9FFDB0]< http://www.ca.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/ (Embedded image moved to file: pic04524.gif) -- 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: ABENDS
> Dan Horák > Sent by: Linux on 390 Port > > one more idea - the address looks as a wrap over kernel memory pages > when there is no page allocated for 4A7A2ED000, but a page is there for > 4A7A2EC000, where the target is being copied, it can be an > "over-optimized" version of memcpy() or something like that > > I already debugged this kind of crash with older glibc in RHEL. Good point. Can this be reproduced with the command running in gdb? If so, what is in /proc//maps at the memory location? cheers, Carsten -- 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: EMC request for GTFTRACE on z/VM and zLinux
Chris, I don't think that this request makes sense. With V=V passthrough, z/VM is'nt involved in data transfer at all. I think that tracing needs to be done in Linux. with kind regards Carsten Otte System z firmware development / Boeblingen lab --- Every revolution was first a thought in one man's mind; and when the same thought occurs to another man, it is the key to that era. - Ralph Waldo Emerson, Essays: First Series, 1841 "Will, Chris" Sent by: Linux on To 390 Port LINUX-390@vm.marist.edu, Subject EMC request for GTFTRACE on z/VM 04.02.2014 15:15 and zLinux Please respond to Linux on 390 Port We are currently having issues with a db2 server running under SLES 11 SP2 using zfcp to access an EMC SAN. EMC has requested a GTFTRACE trace. Is there even a way to run GTF under z/VM and would it make any sense since this seems to be a Linux issue? If there is a way, what is the process to run a GTF trace (I would assume it would have to be run under maint or some other z/VM support ID). Chris Will Systems Software (313) 549-9729 cw...@bcbsm.com The information contained in this communication is highly confidential and is intended solely for the use of the individual(s) to whom this communication is directed. If you are not the intended recipient, you are hereby notified that any viewing, copying, disclosure or distribution of this information is prohibited. Please notify the sender, by electronic mail or telephone, of any unintended receipt and delete the original message without making any copies. Blue Cross Blue Shield of Michigan and Blue Care Network of Michigan are nonprofit corporations and independent licensees of the Blue Cross and Blue Shield Association. -- 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: ABENDS
According to the progam check table in my Principles of Operation reference summary, interruption code (= program check code) 0x60004 is a protection exception. The program is writing to memory that is memory mapped as read-only. This is a user program, therefore the next step is to create a core dump and see why it did that. with kind regards Carsten Otte System z firmware development / Boeblingen lab --- Every revolution was first a thought in one man's mind; and when the same thought occurs to another man, it is the key to that era. - Ralph Waldo Emerson, Essays: First Series, 1841 Tom Huegel To Sent by: Linux on LINUX-390@vm.marist.edu, 390 Port cc Subject ABENDS 04.02.2014 00:07 Please respond to Linux on 390 Port I'm a LINIX dummy. What does this mean? I just installed s390x FEDORA 20 on a z196. [ 5372.889929] User process fault: interruption code 0x60004 in libc-2.18.so [4a7a25+1ae000] [ 5372.889936] failing address: 4A7A2ED000 [ 5372.889953] CPU: 0 PID: 45008 Comm: getent Not tainted 3.12.8-300.fc20.s390x #1 [ 5372.889956] task: 3dc208a8 ti: 13aa8000 task.ti: 13aa8000 [ 5372.889964] User PSW : 070520018000 004a7a2890ee (0x4a7a2890ee) [ 5372.889967]R:0 T:1 IO:1 EX:1 Key:0 M:1 W:0 P:1 AS:0 CC:2 PM:0 EA:3 User GPRS: 004a7a3ca8a2 004a7a2edfe1 004a7a3ca32e [ 5372.889987]0011 03fff6cd7bdc 004a7a404000 03a89f4d [ 5372.889992]03a89f4d 03a88d50 03a88cd8 [ 5372.889996]004a7a402000 004a7a3bd328 004a7a2890e2 03a88cb0 [ 5372.890006] User Code: 004a7a2890dc: c0e5000324b0brasl %r14,4a7a2eda3c 004a7a2890e2: c01a0be0 larl%r1,4a7a3ca8a2 #004a7a2890e8: c03a0923 larl%r3,4a7a3ca32e >004a7a2890ee: d20d20001000 mvc 0(14,%r2),0(%r1) 004a7a2890f4: b904002a lgr %r2,%r10 004a7a2890f8: c0e500021fe0 brasl %r14,4a7a2cd0b8 004a7a2890fe: b9020062 ltgr%r6,%r2 004a7a289102: a78401c8 brc 8,4a7a289492 [ 5372.890105] Last Breaking-Event-Address: [ 5372.890110] [<004a7a2eda5c>] 0x4a7a2eda5c [ 5372.891317] User process fault: interruption code 0x60004 in libc-2.18.so [4a7a25+1ae000] [ 5372.891325] failing address: 4A7A2ED000 [ 5372.891330] CPU: 0 PID: 45006 Comm: getent Not tainted 3.12.8-300.fc20.s390x #1 [ 5372.891335] task: 3d7788a8 ti: 13a7 task.ti: 13a7 [ 5372.891343] User PSW : 070520018000 004a7a2890ee (0x4a7a2890ee) [ 5372.891348]R:0 T:1 IO:1 EX:1 Key:0 M:1 W:0 P:1 AS:0 CC:2 PM:0 EA:3 User GPRS: 004a7a3ca8a2 004a7a2edfe1 004a7a3ca32e [ 5372.891360]0011 03fff6e02bdc 004a7a404000 03984f4d [ 5372.891432]03984f4d 039842b0 03984238 [ 5372.891435]004a7a402000 004a7a3bd328 004a7a2890e2 03984210 [ 5372.891441] User Code: 004a7a2890dc: c0e5000324b0brasl %r14,4a7a2eda3c 004a7a2890e2: c01a0be0 larl%r1,4a7a3ca8a2 #004a7a2890e8: c03a0923 larl%r3,4a7a3ca32e >004a7a2890ee: d20d20001000 mvc 0(14,%r2),0(%r1) 004a7a2890f4: b904002a lgr %r2,%r10 004a7a2890f8: c0e500021fe0 brasl %r14,4a7a2cd0b8 004a7a2890fe: b9020062 ltgr%r6,%r2 004a7a289102: a78401c8 brc 8,4a7a289492 [ 5372.891457] Last Breaking-Event-Address: [ 5372.891460] [<004a7a2eda5c>] 0x4a7a2eda5c [ 5372.891556] Pid 45006(getent) over core_pipe_limit [ 5372.891558] Skipping core dump -- 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
Re: TX-Errs on hipersocket interface.
Alan wrote: >Carsten, this is information that really needs to be in the Device Driver >book, as it differs from the traditional interpretation of TX/RX counters. Ah, you're right - that would be accouned as dropped, not TX error. Forget about my reply then, time to look at other causes for that TX error. Sorry for the noise. with kind regards Carsten Otte System z firmware development / Boeblingen lab --- Every revolution was first a thought in one man's mind; and when the same thought occurs to another man, it is the key to that era. - Ralph Waldo Emerson, Essays: First Series, 1841 -- 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: TX-Errs on hipersocket interface.
Ron, TX errors on a Hipersocket are business as usual. When the receiver fails to pick up packets at the same rate that the transmitter transmits, the queues will run full over time. TCP uses packet loss as a measure to tune the TCP window size, which effectively will reduce the transmit rate. Thus, the packet loss you see only indicates that trimming down transmits to match the receiver's performance is taking place. with kind regards Carsten Otte System z firmware development / Boeblingen lab --- Every revolution was first a thought in one man's mind; and when the same thought occurs to another man, it is the key to that era. - Ralph Waldo Emerson, Essays: First Series, 1841 Ron Foster To Sent by: Linux on LINUX-390@vm.marist.edu, 390 Port cc Subject TX-Errs on hipersocket interface. 24.07.2013 22:13 Please respond to Linux on 390 Port Hello, While investigating another problem, I did a netstat -I on our production SAP Linux App Servers. To my surprise, I am getting TX-ERRS on a hipersocket interface on two of our systems. The two systems have recently been converted to SLES11 SP2. The hipersocket device is connected to a z/OS system where our DB2 system lives. The hipersocket device is also used by other SAP APP servers that are SLES10 SP4. Those servers have zeros in the TX-ERRS column. Any ideas on how to trouble shoot Transmit errors on a hipersocket? Ron Foster Baldor Electric Company 5711 R S Boreham Jr Street Fort Smith, AR 72901 Phone:479-648-5865 Fax:479-646-5440 Email: ron.fos...@baldor.abb.com<mailto:ron.fos...@baldor.abb.com> IM Address:rfos...@baldor.com www.baldor.com<http://www.baldor.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/ -- 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: Who's using execute in place?
Hi Florian, please note that the DCSS block device and its use for paging and/or storing file systems on it is not at risk of being discontinued. Are any of your filesystems ext2 with -o xip? with kind regards Carsten Otte System z firmware development / Boeblingen lab --- Every revolution was first a thought in one man's mind; and when the same thought occurs to another man, it is the key to that era. - Ralph Waldo Emerson, Essays: First Series, 1841 Florian Bilek To Sent by: Linux on LINUX-390@vm.marist.edu, 390 Port cc Subject Re: Who's using execute in place? 12.02.2013 16:48 Please respond to Linux on 390 Port Dear Carsten, We are using DCSS as replacement of a paging device on physical DASD. Works great. I also run the production environment from a DCSS with root and /usr filesystem loaded. So I consider it as a very valuable facility. Kind regards, Florian On Tue, Feb 12, 2013 at 10:36 AM, Carsten Otte wrote: > Dear Linux on z community, > > a few years ago we've introduced execute in place, which can be used to > save some memory by using z/VM DCSS segments. Since the size of > main memory for virtual servers has increased much faster than the size > of binary executables and libraries since, this technique has become less > attractive. Andrew Morton, the leading maintainer for the memory manager > in Linux, has raised the question if this is still needed. > > Who's using execute in place in their environment today? What are your > plans of future use? Can we discontinue the technology or shall we keep > it around? > > with kind regards > Carsten Otte > System z firmware development / Boeblingen lab > --- > Every revolution was first a thought in one man's mind; > and when the same thought occurs to another man, it is the key to that era. > > - Ralph Waldo Emerson, Essays: First Series, 1841 > > -- > 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/
Who's using execute in place?
Dear Linux on z community, a few years ago we've introduced execute in place, which can be used to save some memory by using z/VM DCSS segments. Since the size of main memory for virtual servers has increased much faster than the size of binary executables and libraries since, this technique has become less attractive. Andrew Morton, the leading maintainer for the memory manager in Linux, has raised the question if this is still needed. Who's using execute in place in their environment today? What are your plans of future use? Can we discontinue the technology or shall we keep it around? with kind regards Carsten Otte System z firmware development / Boeblingen lab --- Every revolution was first a thought in one man's mind; and when the same thought occurs to another man, it is the key to that era. - Ralph Waldo Emerson, Essays: First Series, 1841 -- 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/
Antwort: Re: KVM on IBM System z
Your qemu process should have kvm file descriptors open in /proc/fds/ and you should see a debug area in /sys/kernel/debug/s390dbf/kvm-/ with kind regards Carsten Otte System z firmware development / Boeblingen lab --- Every revolution was first a thought in one man's mind; and when the same thought occurs to another man, it is the key to that era. - Ralph Waldo Emerson, Essays: First Series, 1841 -Linux on 390 Port schrieb: - An: LINUX-390@vm.marist.edu Von: Tobias Doerkes Gesendet von: Linux on 390 Port Datum: 11.10.2012 07:15 Betreff: Re: KVM on IBM System z Hi all, one more question regarding KVM on IBM System z: Is there a way to check wether KVM is using hardware virtualisation (SIE instruction)? I installed SLES 11 and virt-host-validate is missing. In FC 17 it returns only software virtualisation: QEMU: Checking for hardware virtualization : WARN (Only emulated CPUs are available, performance will be significantly limited) QEMU: Checking for device /dev/vhost-net : PASS QEMU: Checking for device /dev/net/tun : PASS LXC: Checking for Linux >= 2.6.26 : PASS But i think virt-host-validate in FC17 has no support for s390x. So i want to check wether SIE is used or not. Kind regards, Tobias. -- 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: tar with tape drives
Hi Tom, if I recall correctly, we reserve the tape on open, and free it on closing the file descriptor. You shout be able to find out which process is using it via "fuser /dev/ntibmX". with kind regards Carsten Otte IBM Linux Technology Center / Boeblingen lab -- omnis enim res, quae dando non deficit, dum habetur et non datur, nondum habetur, quomodo habenda est -- 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: SwapCache
Shane: My understanding is that the swap-cache is merely a different "state" Shane: of pages in the page-cache. That's correct. Anonymous pages (=those that do not have a file backing e.g. belong to an executable file or shared library) that are considered not frequently used get a swap slot assigned before paging them out. At this point in time they are considered swap cache. Now the page gets written out and eventually removed from memory due to demand for other content. Once the page gets accessed again, we read it and consider it swap cache: if it needs to be removed from memory again, but has not changed since reading from disk, we can save the I/O to write it out and just recycle the memory page. Once swap space is 50% full Linux frees swap slots on paging in. In this case, those pages are considered anonymous again (not swap cache). By doing so, Linux avoids the potential problem that swap space runs full by potentially having more page-out I/O to the swap target (performance degradation). with kind regards Carsten Otte IBM Linux Technology Center / Boeblingen lab -- omnis enim res, quae dando non deficit, dum habetur et non datur, nondum habetur, quomodo habenda est -- 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/
Fw: SwapCache
Actually I found out that my answer does not answer your question: Linux does never allocate two copies of the same memory page. Any page that is considered as swap cache is in active use by an application and becomes an anonymous page on swapoff. This raises the question why one would care about the size of swap cache in the first place. with kind regards Carsten Otte IBM Linux Technology Center / Boeblingen lab -- omnis enim res, quae dando non deficit, dum habetur et non datur, nondum habetur, quomodo habenda est -- 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: SwapCache
Hi Dave, the kernel will automagically take care of that double allocation: as soon as your swap space is 50% full, the kernel will free the swap slot on paging in. Note that this does'nt save you a single memory page though, since the very same memory content will now be billed as anonymous memory rather than swap cache. with kind regards Carsten Otte IBM Linux Technology Center / Boeblingen lab -- omnis enim res, quae dando non deficit, dum habetur et non datur, nondum habetur, quomodo habenda est Dave Czajkowski To Sent by: Linux on LINUX-390@vm.marist.edu 390 Port cc Subject SwapCache 02.06.2011 14:43 Please respond to Linux on 390 Port Is there a way to disable or limit the amount of SwapCache? I understand it's function, but the double allocation of these pages is causing my low memory alerts to trip. I can remove the swap copy with a swapoff/swapon. Is there a more elegant way to remove? -- 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: z/VM bug, Linux bug, or operator error?
Hi Mark, good point, I am going to address the documentation side. with kind regards Carsten Otte IBM Linux Technology Center / Boeblingen lab -- omnis enim res, quae dando non deficit, dum habetur et non datur, nondum habetur, quomodo habenda est Mark Post To Sent by: Linux on LINUX-390@vm.marist.edu 390 Port cc Subject Re: z/VM bug, Linux bug, or operator error? 27.05.2011 16:17 Please respond to Linux on 390 Port >>> On 5/26/2011 at 05:14 AM, Carsten Otte wrote: > Hi Mark, > > could you please trace diag x64 and send the output? You see this message > due to a diag call failing, > and the trace would enable the hypervisor people to comment on why that is. Hi, Carsten, Thanks for the offer to help, but the fix was Kris Buelens' comment that a DCSS must start and end on megabyte boundaries. That means the "end page" value should end in x'FF', or at least the subsequent DCSS needs to have a "start page" value ending in x'00' to avoid overlaps. That fact should probably explicitly be called out in the XIP documentation. (In a blatant attempt to deflect blame for my shallow z/VM skills. :) 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: z/VM bug, Linux bug, or operator error?
Hi Mark, could you please trace diag x64 and send the output? You see this message due to a diag call failing, and the trace would enable the hypervisor people to comment on why that is. with kind regards Carsten Otte IBM Linux Technology Center / Boeblingen lab -- omnis enim res, quae dando non deficit, dum habetur et non datur, nondum habetur, quomodo habenda est Mark Post To Sent by: Linux on LINUX-390@vm.marist.edu 390 Port cc Subject z/VM bug, Linux bug, or operator error? 25.05.2011 02:49 Please respond to Linux on 390 Port Cross-posted to IBMVM and Linux-390 I'm working on a project where I want an xip2 file system that is larger than 2GB. So, I'm trying to use the support built into z/VM and Linux to logically concatenate multiple DCSS into one. But, I can't seem to get it to work. I can load one segment or the other, but not both. The message that comes out on the console is extmem.cb0afe: ITMV1B needs used memory resources and cannot be loaded or queried "man extmem.cb0afe" says "You cannot load or query the DCSS because it overlaps with an already loaded DCSS or with the memory of the z/VM guest virtual machine (guest storage)." I don't believe any of those apply, but I need someone to tell me what to do differently, or to whom I should be complaining. The segments are defined as this: FILE FILENAME FILETYPE BEGPAG ENDPAG TYPE CL #USERS 0054 ITMV1A DCSSG00028 0002FFE00 SR A 0 0055 ITMV1B DCSSG0002FFE01 00037FD00 SR A 0 The guest virtual storage is 1GB: #CP Q V STOR STORAGE = 1G The kernel was booted with a mem= value sufficient to contain the highest address of the ITMV1B DCSS: # cat /proc/cmdline root=/dev/disk/by-path/ccw-0.0.9300-part1 TERM=dumb mem=16384M BOOT_IMAGE=0 The command to allocate the two DCSS is: echo itmv1a:itmv1b > /sys/devices/dcssblk/add And that's where I get my error (after 2 minutes). As I said previously, if I try to load them individually, I can do one or the other, but not both. This is an up to date SLES11 SP1 system, running on z/VM 6.1 RSU 1003. Does anyone see the problem? 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: GPG Key Ring Generation on zLinux Fails
Hi Mark, on the mainframe entropy is not so easy as there's no computer mouse to move around ;-). I recommend to do some I/O intensive work to a FICON disk, the device driver is our prominent source of entropy (moving platters seek times least significant bits, maybe we need to rethink that in the light of upcoming SSDs) . Oh, and the crypto cards do have a hardware entropy source, but I don't know if GPG would use that. with kind regards Carsten Otte IBM Linux Technology Center / Boeblingen lab -- omnis enim res, quae dando non deficit, dum habetur et non datur, nondum habetur, quomodo habenda est Mark Jacobs To Sent by: Linux on LINUX-390@vm.marist.edu 390 Port cc Subject GPG Key Ring Generation on zLinux Fails 20.05.2011 16:00 Please respond to mark.jacobs@custs erv.com I'm attempting to generate a key ring in a zLinux environment using gpg but I can't get enough entropy to supply the generation process with enough random bytes. Not enough random bytes available. Please do some other work to give the OS a chance to collect more entropy! (Need 284 more bytes) I've tried everything I can think of, and my zLinux support team says that this is a known problem with virtualized environments. Does anyone have any suggestions on how to get key ring generation to reliably work on zLinux? -- Mark Jacobs Time Customer Service Tampa, FL Some people are electrifying, they light up a room when they leave. -- 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: zLinux kernel architecture
On 27.04.2011 16:22, Alan Cox wrote: Well perhaps you should get it from the proper source ? http://oreilly.com/catalog/9780596005658 Yea, that's the link that I was trying to find via google.de. Thanks, Alan :-). -- 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: zLinux kernel architecture
I would recommend http://www.ebook3600.com/understanding-the-linux-kernel-third-edition It does'nt have any z specifics, but it gives a very informative, fun reading overview about the kernel. with kind regards Carsten Otte IBM Linux Technology Center / Boeblingen lab -- omnis enim res, quae dando non deficit, dum habetur et non datur, nondum habetur, quomodo habenda est -- 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: z/VM RedHat Virtual Machine Memory abend
This is a kernel bug. If this is a GA kernel, open a service request against RH/IBM to get it fixed. with kind regards Carsten Otte IBM Linux Technology Center / Boeblingen lab -- omnis enim res, quae dando non deficit, dum habetur et non datur, nondum habetur, quomodo habenda est -- 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: New User Linux on z10 question
> While 12 LPARs can share one IFL they cannot share any memory. Each LPAR must have its own dedicated memory. That's the main reason to run z/VM, to share memory. When > I asked why Xen did not support System z, I was told that they didn't see any point in competing with z/VM. Xen is architecture dependend code all the way through. I know it also runs on Power, but the only commonality between Power and x86 is probably the printk routine... with kind regards Carsten Otte IBM Linux Technology Center / Boeblingen lab -- omnis enim res, quae dando non deficit, dum habetur et non datur, nondum habetur, quomodo habenda est -- 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: read-only
> We are using multipath daemon Carsten... The configuration looks good to me. And you're mounting /dev/mapper/*, and you see those filesystems remounted readonly after recovery? I sense a great disturbance in the Force. If that is so, I think it's about time to think about opening a service request. with kind regards Carsten Otte IBM Linux Technology Center / Boeblingen lab -- omnis enim res, quae dando non deficit, dum habetur et non datur, nondum habetur, quomodo habenda est -- 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: read-only
Your message does'nt tell me anything about your setup. I'm guessing that you're not using the SCSI multipath daemon and use sdX devices directly. In that case I think you should a) get a second path to your device, b) setup multipath daemon, and c) use queue_if_no_path setting to ensure that your filesystems remain usable. with kind regards Carsten Otte IBM Linux Technology Center / Boeblingen lab -- omnis enim res, quae dando non deficit, dum habetur et non datur, nondum habetur, quomodo habenda est -- 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: New User Linux on z10 question
> Someone told me it is possible to run 12 Linux images on a single IFL > without the use of z/VM. > Is this true and how? One thing you could do is define 12 logical partitions (LPAR). You can define them IFL only, with shared engines. Not that I'm advising to do so with kind regards Carsten Otte IBM Linux Technology Center / Boeblingen lab -- omnis enim res, quae dando non deficit, dum habetur et non datur, nondum habetur, quomodo habenda est -- 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: cross compiling for s390
Last time I was on decaf and had to compile on x86, I used that and back then it worked fine for me: http://debian.speedblue.org/ Note that mainframes tend to have more CPU cores than laptops, which is good for compiler performance when using make -j ;-) with kind regards Carsten Otte IBM Linux Technology Center / Boeblingen lab -- omnis enim res, quae dando non deficit, dum habetur et non datur, nondum habetur, quomodo habenda est adam shea To Sent by: Linux on LINUX-390@vm.marist.edu 390 Port cc Subject cross compiling for s390 04.04.2011 21:01 Please respond to Linux on 390 Port I would like to set up gcc on my laptop to cross compile for s390 architecture. Has anyone attempted this? Pointers and words of wisdom would be appreciated greatly. -ajs -- "Intelligence is the ability to avoid doing work, yet getting the work done." -Linus Torvalds -- 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: Where is kernel loaded in memory?
Quote: The larger problem is going to be getting buy-in from the app owner to implement such a thing. Actually, all you need to do is _install_ it on subject file system. There's nothing on the implementation side that needs to be done for it. with kind regards Carsten Otte IBM Linux Technology Center / Boeblingen lab -- omnis enim res, quae dando non deficit, dum habetur et non datur, nondum habetur, quomodo habenda est Mark Wheeler To Sent by: Linux on LINUX-390@vm.marist.edu 390 Port cc Subject Re: Where is kernel loaded in memory? 18.03.2011 13:29 Please respond to Linux on 390 Port Mark, That's an excellent suggestion, inasmuch as we have a couple dozen of these servers running and, if nothing else, sharing code could reduce overall demand for memory. Obviously, the bigger win would be if this is the code that is actually being paged in and causing the delays, since by sharing it would be more likely to be in storage already. The larger problem is going to be getting buy-in from the app owner to implement such a thing. Best regards, Mark > Date: Thu, 17 Mar 2011 14:55:07 -0600 > From: mp...@novell.com > Subject: Re: Where is kernel loaded in memory? > To: LINUX-390@VM.MARIST.EDU > > >>> On 3/17/2011 at 04:47 PM, Marcy Cortes wrote: > > When you have some memory intensive night time activities, like say scanning > > checks in cron jobs, or heavy batch stuff, these things happen. We added > > more memory :(, although lowering the heap size did help some. > > > > Anyway you can prime the pump (script to touch your pages in a cron job?)? > > Putting the application into an xip2 file system on a DCSS might help with this quite a bit. > > > 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/ -- 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: Where is kernel loaded in memory?
Quote: I note our German maintainers have been conspicuously quiet Allright, so here's the short answer to the original question. You can build the kernel with either a tape or a card reader boot loader. When you IPL that, it gets loaded into guest real memory starting from address 0. However, if you use zipl to prepare a disk, the boot code will skip the first 64k and start loading the rest of the kernel above 64k. When the kernel starts up, it will allocate addional memory for its purposes such as memory management or task management data structures, and also for loadable kernel modules. At the same time, it will free up memory during startup like kernel functions that are only used for initialization (kernel message: "freeing bootmem"). The resulting footprint for both kernel binary code and its data structures will be spread over the guest real address space, with a big chunk being the kernel text section starting at 64k ending at about 5meg depending on your kernel configuration. Since this guest real address space is seperate from both the application address spaces and the LPAR/VM hypervisor address spaces it does not matter to either an application and/or a hypervisor where the kernel and its data structures are located in memory. with kind regards Carsten Otte IBM Linux Technology Center / Boeblingen lab -- omnis enim res, quae dando non deficit, dum habetur et non datur, nondum habetur, quomodo habenda est -- 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: ECKD driver vs DIAG driver
Quote: > Mark is correct: one automagically created partition. Worse, there > is no 'fdasd' or 'fdisk' management of that partition. WORSE STILL, > you *must* put the filesystem into the "partition" (such as it is) if > you are going to boot from this disk. A filesystem in /dev/dasdx will > be clobbered by the first stage of the boot loader, while a filesystem > in /dev/dasdx1 is protected by the extra 8K of padding. (12K total) > I checked it again this morning. The bootstrap overwrites the root inode. Well the LDL disk layout basically consists of two blocks of data (size depends on the blocksize used for formatting) Those are being used to a) label the device and b) conatin the IPL boot code (channel programs). In theory, we could mangle the IPL code into the first 512 bytes of a filesystem which is reserved on x86 as "partition boot record", but that'd have a couple of downsides: a) the disk would not be "labled" as observed by other mainframe OSes b) this would differ from how people use Linux on other platforms where they typically do use partitions in front of the partition. Note that for LDL formatted media, you may chose to put the filesystem on the device itself (like /dev/dasdx instead of /dev/dasdx1) for volumes that you do not intend to boot from. Due to the fact that for ECKD CDL media blocks on track 0 do not have the formatted size, you cannot do the same with ECKD CDL formatted disks (filesystem corruption would be the result). with kind regards Carsten Otte IBM Linux Technology Center / Boeblingen lab -- omnis enim res, quae dando non deficit, dum habetur et non datur, nondum habetur, quomodo habenda est -- 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: Mounting a filesystem *really* read-only
Am Sun, 14 Jun 2009 01:34:35 +0200 schrieb Rob van der Heij : > One of the problems is that S/390 has no way to sense the R/O setting > of the disk, so we can't pass that to the mount (like you have with a > CD for example). The only way to tell is by trying to write to the > disk.We did not think that was very elegant. > > But you can also tell the Linux dasd driver to not write it, that > should be more powerful than mount ro. It's probably through the > hwcfg-* files. I believe that with that setting the mount will default > to "ro" ? You recall correctly. The parameter can also be set via the (ro) suffix on dasd parameter line either as parameter to insmod when loading dasd, or via kernel parameter (zipl). Like this: dasd=1000,1001(ro),2000-20ff(ro) ^^ dasd 1000 will be writable, 1001 and 2000-20ff are read only. All mounts of subject dasds will be read-only by default. The parameter can be changed using the blockdev command (--setro option)*. * make sure you're not setting a disk to read-only that is currently mounted read+write -- 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: Mounting a filesystem *really* read-only
Am Fri, 5 Jun 2009 16:47:15 -0700 schrieb Ted Rodriguez-Bell : > Now I think we could work around this by changing the device to > read-only before mounting, so you'd do something like this: > blockdev --setro /dev/dasdXX # Lock the device > if mount /dev/dasdXX /mount/point ; then > if [[ ! -f /mount/point/IN_USE ]]; then > SAFE_TO_MOUNT=yes > else > SAFE_TO_MOUNT='no, lock file exists' > fi > else > SAFE_TO_MOUNT='no, could not mount filesystem' > fi > blockdev --setrw /dev/dasdXX > I've tested "blockdev --setro" and it *is* local; changing it on one > machine doesn't change the filesystem on the other. > > But is there some other reason to think this might be unsafe? Would > Novell support be a better forum than this one? Using a lock file on the filesystem looks racy to me. Did you consider using the reserve/release feature for dasds for this? In case of unexpected crash, you could break the lock and recover the filesystem once you're sure the system that was holding the lock has died indeed. so long, Carsten -- 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: KVM
Am Thu, 4 Dec 2008 09:59:30 -0500 schrieb Alan Altmark <[EMAIL PROTECTED]>: > I can represent IBM with respect to z/VM: IBM is not getting rid of z/VM > in favor of KVM. > > IBM is fully aware of KVM and it's potential intersection with z/VM should > a System z version ever emerge. It does not in any way devalue z/VM. > > Anyone who says otherwise is ill-informed or simply trying to create FUD. > If you hear such a message from an IBMer, please send me details of the > contact and I will see to it that they receive a proper education. I second that. For those interrested in the current status of KVM on the platform, see http://kvm.qumranet.com/kvmwiki/KvmForum2008?action=AttachFile&do=get&target=kdf2008_17.pdf This presentation is from spring 2008, the stop machine run and memory overcommitment issues have been fixed by now as of 2.6.28-rc. If you want to try it out, see: http://www.ibm.com/developerworks/linux/linux390/kuli.html so long, Carsten -- 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: How to virtualize Windows under SLES Linux on zSeries - PJBR
Am Fri, 14 Nov 2008 22:51:14 +0900 schrieb John Summerfield <[EMAIL PROTECTED]>: > QEMU is different. I don't know what the interaction with KVM is, but > QEMU is said to emulate fairly well. I've not explored setting up QEMU > to run Power, but I did try QEMU on my G4 laptop a while ago. I decided > Windows was too slow to be useful. KVM just uses qemus device subsystem for I/O emulation and it uses it as a management hub. The qemu console allows to attach/detach devices/cpus and such similar to CPs console. so long, Carsten -- 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: Implementing a Shared Root File System
Am Wed, 12 Nov 2008 10:01:39 -0500 schrieb Craig Loubser <[EMAIL PROTECTED]>: > It is SLES10 SP2 (kernel 2.6.16.60-0.21). SLES10 SP2 is definitely > behaving differently to SLES9. Under SLES9, prior to passing control to > /sbin/init, it would mount the root file system as R/O and this is > without marking the boot device as "ro" in the kernel command line. Right. The behavior of Sles9 is as expected, and as documented in our device driver book. > Since we have a support contract with > IBM for Linux on z, I will raise a PMR just to see what they say. Once > again, thanks to everyone on this list. Ok good, I've already given the dasd driver owner an heads-up. so long, Carsten -- 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: Implementing a Shared Root File System
Am Mon, 10 Nov 2008 19:16:19 -0700 schrieb Mark Post <[EMAIL PROTECTED]>: The fsck is not what is causing the I/O errors. When you see "fsck succeeded" it's done and over with. Let me guess, though. Is this an ext3 file system (or reiserfs), and are you mounting it as ext3 instead of ext2? If so, then I suspect the system is trying to replay the journal, resulting in the I/O errors. There's no point in having a journaling file system that is only going to be mounted read-only. Use ext2 instead. Neither should the file system nor fsck be able to write to the volume in the first place, if it is marked (ro). Also, it should be mounted read-only by default even if you forget the additional ro in the kernel command line as long as the device itself is read-only - and that should be the case here, due to the dasd=100(ro) in the kernel command line. This clearly is a kernel bug, Craig should open a PMR for that. What distribution/kernel version are you running? so long, Carsten -- 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: Implementing a Shared Root File System
Am Fri, 7 Nov 2008 15:46:56 -0500 schrieb Craig Loubser <[EMAIL PROTECTED]>: > Waiting for device /dev/dasda1 to appear: ok > rootfs: major=94 minor=1 devn=24065 > fsck 1.38 (30-Jun-2005) > Ý/bin/fsck.ext2 (1) -- /¨ fsck.ext2 -a /dev/dasda1 > /dev/dasda1: clean, 51899/874368 files, 299859/1748676 blocks > fsck succeeded. Mounting root device read-write. > Mounting root /dev/dasda1 > end_request: I/O error, dev dasda, sector 192 The disk backing dasda1 is apparently read-only in VM, but has not been marked ro in linux. Try dasd=devno(ro) for this volume in your kernel command line. -- 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: OCSF2 support
Am Wed, 29 Oct 2008 15:35:34 -0500 schrieb Marcy Cortes <[EMAIL PROTECTED]>: > Still looking for a clustered file system (not a networked file system) > that will allow more than one server to r/w to the same disk. > Several applications have wanted this. (We use Veritas VCS on > distributed but that does not run on z ). OCFS has grown mature on x86, but be aware that not too many people use it on s390 so far. The code looks clean and nice, and thus I'd not expect a lot of problems in there. You should give it an extended testing in your environment before betting your job on it, but time has indeed come to use it for business. so long, Carsten -- 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: HSM-like function on Linux on z
Am Mon, 20 Oct 2008 16:01:16 -0600 schrieb Lee Stewart <[EMAIL PROTECTED]>: > Over a year ago there was a discussion with people looking for an > HSM-like migrate and retrieve function for Linux. Back then it sounded > like there was interest, but no solutions. Since it's been a while, > has anyone heard of anything new in this space? As far as I know, only XFS has support for it in the kernel. All other file systems don't and I don't know of a user space application that makes use of XFS'es feature. It would be an interresting use-case for fuse (http://fuse.sourceforge.net/) though, but I have'nt heared of anyone working on an implementation for that (and none is listed in the fuse file systems list). so long, Carsten -- 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: Root filesystem error switches to ReadOnly
Rob van der Heij wrote: I expect the problem is using 'dd' for the copying may not get the magic signatures that makes the disk a CDL format, so the driver ends up seeing it as LDL and gets things misaligned. But you should be able to notice that when the new system is booting. This is not the case. Using dd on the entire disk does copy all data in the "magic" first tracks, and blockdev --rereadpt /dev/target/disk will cause the kernel to reread it on the fly. Just make sure the target is really the same disk layout as the source upfront, because otherwise you'll end up in deep because the dasd driver's media detection will see a different disk layout than the partition detection code. -- 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: Root filesystem error switches to ReadOnly
Fargusson.Alan wrote: Doing a dd on a device goes through the cache, so in this case the cache can't be the problem. This is not the case. In fact, the page cache is indexed by address spaces and each file as well as each block device has its own address space. Consequently, Linux cannot tell that you're reading/writing the same block on disk via the device node that is already in the cache because it belongs to a file of a mounted filesystem that relies on there. The only safe way of doing an online snapshot of a mounted file system is the dm-snapshot target. Flashcopy and DD don't do the trick. -- 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: swap to DCSS documentation/cookbook?
Pieter Harder wrote: I am starting to think about the possibilities 2GB above the bar DCSS in z/VM 5.4 will provide for replacing large Vdisk swap. But I can't find any usable documentation on how to do swap to DCSS. The basics are documented on dw and I can probably manage, but when anyone has been there and done that can you please post some pointers and save us all the trouble of finding out again. Performance information is quite welcome as well. The dcssblk device driver is a much faster swap target then vdisk. We have'nt got a good documentation on the setup unfortunatly, but I am willing to guide you through the setup process. In short, you'll have to do the following steps: On _one_ privclass E guest do: - use DERFSEG to define a dcss with one page "exclusive write" and the rest "exclusive nonsaved" access mode - define the guest storage larger then the dcss, then ipl CMS and save the empty segment - reduce the guest storage again, boot linux, load the dcss in the dcssblk driver (should report EW/EN mixed mode segment in dmesg) - do mkswap on the block device, and save the segment again via the "save" attribute of the dcssblk device driver On every guest do: - add "dcssblk.segments=NAME" to your kernel parameter line to get the block device active on startup, and add the swap target to /etc/fstab like this: /dev/dcssblk0 swapswappri=1 0 0 - do swapon -a, and verify that the swap is active in /proc/swaps -- 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: Redhat 5.2 zipl default configuration problem
Patrick Spinler wrote: I'm experiencing an issue with zipl on redhat 5.2, where it doesn't appear to create a usable default for it's boot menu. Has anyone seen anything similar? I just opened a PMR with IBM, but thought I'd ask here, also. Interresting. When reading the config file, it looks to me like "linux" is the default, and that boot number 0 (default) should be the same target as boot number 1 (linux). Your zipl boot menu output on IPL should also state "default (linux)" instead of "default (-)". I'd say this looks like a bug in the bootloader. RedHat does patch our s390-tools to their s390utils, therefore this bug might either be in our s390-tools also, or it is introduced by redhats patches that go on top. Opening a PMR for it seems reasonable to me. so long, Carsten -- 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: question on s390-tools
LJ Mace wrote: I've downloaded the most recent tools package and have unzipped it. I looked through the readme file but didn't find the install instructions. I have gotten use to yast and also have done the rpm thing(thanks to Mark Post) but I am unsure with this product. What do I need to do to install this package After unzipping, changing into the s390tools subdirectory and typing a "make install" as root user should do the trick. so long, Carsten -- 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: Reserved blocks on ext2/ext3 filesystems?
Ted Rodriguez-Bell wrote: Has anyone done any testing to see what the right reserved block percentage (-m to tune2fs or mke2fs) is? On the Berkeley Fast Filesystem the right number was 10%; the ext2/ext3 default is 5% so I assume that's pretty good for most Linux. But with so many layers between Linux and the hardware, is that a good default on zSeries? I ask because we've just noticed that Levanta left us with filesystems where -m was 0. That gives us an interesting time with filesystem monitoring (if you want an alarm at 90% of capacity, does it trigger at 90% of the total blocks or 85.5%?), but we're also wondering if we have any performance-related time bombs. This is really not a question of the file system technology, but merely depending on the use of the file system. The reserve is _not_ for any file system meta data but only for the root user. The system is supposed to be able to write log files, and allow the root-user to recover from the situation in case the file system runs full. For file systems like /usr, a reserved blocks percentage of 0% is just fine. Same for any data-disks that contain application data. For the disks backing /root and /var, I'd recommend to leave some space reserved there. Also remember that the reserved percentage is relative to your file system size. If you use many small file systems, you probably want more reserved percentage then on a large one. so long, Carsten -- 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: Hw compression of 3592 tape data
Juha Vuori wrote: I understand that there should be a compression feature inside this tape system. According to Device Drivers, Features, and Commands - SC33-8289-04, it could be activated in tape_3590 driver with # mt -f /dev/rtibm0 compression 1 but "mt" in this distribution does not support "compression" command. I've looked at the mt in my fedora installation, and it does'nt have compression as well. You need something that ends up calling this one [from /usr/include/linux/mtio.h]: #define MTCOMPRESSION 32/* control compression with SCSI mode page 15 */ According to gnu.org, gnu mt does'nt support this call at all: http://www.gnu.org/software/tar/manual/html_node/mt.html You need to have the mt-st tool, which is an extended version of mt. Try to install the mt-st rpm that comes with your distribution, it should support --compression as stated above. Looks like we need to update the documentation there. so long, Carsten -- 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: Ext3 (Journaling Filesystems) and BARRIERS = Filesystem Integrity
Mark Perry wrote: Better if this were answered by a z/VM and Linux developer, and hardware development, and., but my 5-cents Your answer is correct. The Linux dasd block device driver will only mark a request as completed (and thus a journal transaction as "done") after the channel program has ended. On Shark, this means that the data has hit the non-volatile storage. Even if you pull the plug now, or any single component fails, everythings safely stored. With synchronous PPRC, this also means that the data is commited to the nonvolatile storage of both storage servers. so long, Carsten -- 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: Ext3 (Journaling Filesystems) and BARRIERS = Filesystem Integrity
Mark Perry wrote: I am not sure how many of the people on this list subscribe to LWN, but there was, for me, an eye-opening article on "Barriers and journaling filesystems": http://lwn.net/Articles/283161/ Thank you Mark, great article. It'll be available public in one week for those who don't subscribe to lwn. -- 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: OCFS2 Setup
Shockley, Gerard C wrote: I'm at the zSeries Oracle Special Interest Group this week and Oracle says "OCSF2 is not supported and non-directional". They say use ASM and RAC. That means the file system is not supported to be used with the database. The file system itself seems pretty solid to me. -- 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: Backup that saves LVM data?
Lee Stewart wrote: Hi All Does anyone know of a backup solution for Linux (on Z) which: a) can be done while the system is up b) also saves the LVM and filesystem metadata (as opposed to just files) Consider using the snapshot target in the device mapper. http://tldp.org/HOWTO/LVM-HOWTO/snapshots_backup.html -- 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: ext2 .v. ext3 (was Re: DASD error on zlinux ipl
Ian S. Worthington wrote: ext2 .v. ext3: I choose ext2 as the system I'm using is slow: glacially so. What's the performance cost in going to ext3? It's not much. Think about using an external journal to a dedicated model1 dasd to speed things up a little. If performance is a strong concern, adventurous minds may also consider to journal to a vdisk. In case VM keeps running, the journal should be okay. In the worst-case, you'll have to run a file system check as with second extended and recreate your journal. cheers, Carsten -- 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: modprobe.conf
Fuhrmann Anna wrote: could you tell me when /etc/modprobe.conf is used - besides when 1. booting the system 2. willingly submitting a modprobe command? Hi Anna, it's used by other means. It contains aliases like "alias char-major-x moduleX". And whenever a device node is accessed from userland that does not have a driver registered for it as described in /proc/devices, the kernel hotplug mechanism will call a usermode helper that will try to resolve that by loading the appropiate device driver automagically after parsing that file. so long, Carsten -- 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: Using xfs file sys
Dott, Robert wrote: We have tried both with and without the -s s=4096, and when we went to mount it, it would not mount. I've seen this too when I tested it back in the old days of 2.4.x. XFS seemed to have trouble with 4k blocksize disks back then, and I reported that it its developers. XFS worked proper when formatting the dasd with 512 bytes blocksize. For that, you have to specify 512 bytes on dasdfmt, and on mkfs. cheers, Carsten -- 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: Kernel BUG at drivers/s390/cio/device_fsm.c:1291
Ron Foster at Baldor-IS wrote: I have a problem that I have not been able to find a solution for, so I have joined the list. Welcome :-). kernel BUG at drivers/s390/cio/device_fsm.c:1291! Anyone have any ideas on what to do or who to contact ? BUG() is a macro that kernel developers use to indicate "something bad went wrong here, and I don't know how to recover". In this case, it is in "our" code, and needs fixing. So, I guess there are two answers: - if you have a service contract, shrink-wrap the kernel's debut output and hand it to your service representative, and make him open a customer problem record. - if not, go the "linux" way and look at /usr/src/linux/drivers/s390/cio/device_fsm.c, contact the developers named in that file and inform them about your situation. They will be interrested to track down the issue on a best-can-do basis. cheers, Carsten -- 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: Linux community, was Re: Demo of OpenSolaris running on Systemz
Dave Jones wrote: But first we would need a really good PL/I compiler, which in itself, is not a bad thing. I would really like to have a reasonable alternative to C for implementing things in Linux, for all of the reasons plus some, that DB mentions below. http://pl1gcc.sourceforge.net/ -- 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: Linux community, was Re: Demo of OpenSolaris running on Systemz
John Summerfield wrote: Don't forget the Windows doesn't actually run very well on Zeds. I'd say "doesn't run," but someone here will stick his hand up and say, "I've done it." Yea, done that in bochs :-) -- 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: z/Series Linux (Debian) I/O Monitoring
Coffin Michael C wrote: Is there a way in z/Series Linux (we're running Debian, FWIW) to monitor disk I/O by userid and/or process? I'm looking for something like "top" but that includes disk I/O statistics for userids/processes. We are seeing a huge increase in disk I/O by our Linux virtual machine guests, but I don't know how to associate this I/O activity with "real" processes or userids to determine if they are legitimate or if there is a problem somewhere. Issue is, that you cannot observe this directly due to the page cache indirection: the block layer does'nt _know_ who's causing disk I/O. On the other hand, you can observe it indirectly: a process that causes heavy disk I/O tends to be in device wait state frequently. You can observe the process state with top. Btw: iostat and vmstat can report the overall I/O rates cheers, Carsten -- 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: high water mark for swap space used?
barton wrote: Wouldn't it be reasonable and mostly trivial that 1) when a process is terminated, that any of it's pages that are on swap are issued a CP diagnose to discard the page? (requires a bit of Endicott work) No. In case a process terminates, the memory is freed. For swap, this does not imply any interaction to the underlying block device. Even if the memory management would care, there currently is no call to do that to a block device. A regular block driver (e.g. dasd) cannot do any reasonable action in case such a call comes up. Bottom line: changes to two different kernel subsystems, cross platform code affected, only useful on s390. Useful but far from being trivial. 2) when a page is swapped in, make that page anonymous again, and issue that same diagnose? This eliminates duplicate resources of expensive storage. This would hurt performance a lot. The trick with sticky swap slots is, that a page that is accessed read only in swap cache always stays clean. Thus, it can be discarded again without the need to page out: we do have an up-to-date swap slot out there. When swapping to regular disks, and thus for 99% of Linux' users, this would be a significant performance degradation because read accesses outnumber write accesses in many pratical scenarios. Even with vdisk, when the guest is memory constrained while CP is not, this could hurt performance overall. cheers, Carsten -- 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: high water mark for swap space used?
Brandt, Mark H wrote: IMHO it would be more efficient to get IBM and Novel to address the issue and eliminate VDISK memory creep without having to configure multiple VDISKS with different priorities. That would fundamentally change the way linux memory management works today: Pages are distinguished in "anonymous" and "non anonymous" ones. Those that are non anonymous have a backing on disk, for example they belong to /bin/bash when running your command interpreter, or to a library. Anonymous ones usually come to life by malloc(). They don't have a backing on disk. When time comes that a page gets to the end of the inactive list, which means it is about to be discarded from main memory, we check if the page is anonymous or not. If is non anonymous (that is, page has a backing on disk) we write back the changes and discard it. For anonymous pages (no disk backing), they get a swap slot assigned from the lowest prio swap available and are considered non anonymous from here on: they're written back to their swap slot and discarded. When that page gets accessed again later, it is considered to be swap cache with a backing on disk: the swap slot. No method is in place to ever change a backing. Thus, a page will never be moved from one swap slot to a different swap slot. This is why linux is not very good at keeping swap usage local, and is not good at using low prio swap for hot pages. btw: I'd like to recommend Mel Gorman's book on this. Pdf is free on the internet. The title is "understanding the linux virtual memory manager". cheers, Carsten -- 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: s390 unit record device driver
Mark Perry wrote: Wouldn't it be nice if Redhat and Suse would backport this, pretty please :-) That would be a reasonable thing to do ;-) Carsten -- 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: Question about using z/VM DCSS for sharing code under Linux
barton wrote: If your system is already tuned with the easy stuff and you already understand how to manage virtual machine sizes, and you use vdisk for swap, then you could save something using XIP. Otherwise, sorry, don't bother for 20 servers. I second that. so long, Carsten -- 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: HSM on Linux on System z
Tim Hare wrote: I'm not that experienced in Linux (System Z or otherwise) - but perhaps this would be a place for a new filesystem, one built from the ground up to handle migration / recall of files in the ABR / HSM manner? In thinking about it, it seems to me that the filesystem code is where you'd handle things like directory entries which don't map to real blocks on disk (i.e. a migrated file), and where you'd handle automatic recall based on someone opening the file. As far as I know, XFS does support this. At least it does on Irix, and I believe this is also true for the Linux implementation. See http://oss.sgi.com/projects/xfs/ , look for "Support for Hierarchical Storage". Whether or not there are HSM solutions available that exploit this filesystem feature is unknown to me. -- Carsten Otte has stopped smoking: Ich habe in 11 Monate, 2 Wochen und 21 Stunden schon 1.679,56 Euro gespart anstatt 6.998,20 Zigaretten zu kaufen. -- 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: OCFS2 and ECKD dasd
Bernard Wu wrote: Can anyone tell me whether OCFS2 works with ECKD dasd devices ? We're running zVM 5.2 and SLES10 . We're still working on getting NPIV to work, so we don't have iSCSI devices to play with. OCFS2 is expected to work with any block device, because it uses the block device abstraction. It does not have any DASD/SCSI/ATA or whatever hardware specifics. Carsten -- 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: Cluster File System for HA
Hi Bernhard, What kind of cluster filesystem is available for SLES9 / SLES10 ? I know there is OCFS for oracle database files, but what about non database files ? actually OCFS2 is meant to be a general purpose cluster file system. The initial version found in Sles 9 was a little early snapshot, which is why it was only promised to work with Oracle databases. The Sles 10 Version however has received a lot of Bug fixes and is accepted upstream. I don't know if Novell officialy supports it for general purpose workload on z, but it is the way best choice available. with kind regards, Carsten -- 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: This Beer's for You
Mark Post wrote: The perfect beer for Linux/390 Systems Administrators http://linuxvm.org/Info/l390beer.html Oh finally you figured our highly confidential design method. Will you promise not to tell anybody? cheers, Carsten -- 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: Alloc memory errors....
Tom Shilson wrote: > The 2-order v. 5-order have to do with the amount of memory requested. The > memory manager pools each have a single request size for efficiency. I > have forgotten the relationship between order number and amount of memory > requested. Very good explanation, Tom. The size = 2 ^ order * PAGE: Order 0 = 2^0 = 1 Page = 4K Order 1 = 2^1 = 2 Pages = 8K Order 2 = 2^2 = 4 Pages = 16K ... You usually only see this message if the requester of the memory sits in kernel space. Note that some kernel components continue normal operation even if their memory allocation fails. Our dasd device driver is a very good example for that. This message therefore does not always indicate an error. -- Carsten Otte has stopped smoking: Ich habe in 10 Monate, 5 Tage und 14 Stunden schon 1.486,05 Euro gespart anstatt 6.191,90 Zigaretten zu kaufen. -- 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: Linux Swap and memory question
Peter E. Abresch Jr. - at Pepco wrote: I have analyzed the LINUXP01 memory usage over last week. I noticed that between the hours of 04:00 – 04:30 AM every morning is when the largest swap increase occurs. 04:15 AM is when some Linux housekeeping starts. These tasks complete in about 3-5 minutes. These tasks require memory and Linux will swap out memory not being used. The memory that is being swapped out is from WebSphere. This is normal Linux process. What is interesting is that this memory is never swapped back in. As the first graph depicts, we see a jump in swap space usage every morning during the housekeeping process. Every day shows another increase in swap space utilization and a corresponding decrease in WebSphere resident memory size about the same time as shown in the second graph. The swap space usage cannot be used as indication whether or not pages are swapped back in: when a page gets swapped out for the first time, it gets a swap slot assigned. This is reflected as swap usage increase. Once it gets swapped back in, it is treated as swap cache: we have the page in memory _and_ on disk. This has two advantages: - if the page does not get changed ("dirty"), we don't need to swap it out again - we can just discard it because we have an up-to-date copy on disk. - if we need to swap it out again, the page does already have a location on disk assigned which removes overhead for housekeeping swap pages when swapping the same page in and out over and over again. If you look at /proc/meminfo the value "SwapCached:" indicates how much was swapped out before and was brought back in. If you substract SwapCached from the swap usage, you get the number you're looking for: how much does reside on disk that is not present in memory. Since the major portion of the swap space utilization is never brought back in, it is obviously not needed. Is this indicative of too much Java heap size which never forces garbage collection? Is there a “memory creep” condition in WebSphere, WebSphere’s JAVA, or one of the applications running under WebSphere? Are the trees blocking my view of the forrest? > > What can I conclude from this? All comments and suggestions are welcome. > The Linux guest size is 768 meg and the Java heap size is min 50 and 700 > max. I don't know about the java internal housekeeping, therefore I cannot do a well-founded statement on this question. I would expect that the jave runtime does page-in all of the memory over time when workload demands it. so long, Carsten -- Carsten Otte has stopped smoking: Ich habe in 10 Monate, 2 Tage und 20 Stunden schon 1.472,80 Euro gespart anstatt 6.136,70 Zigaretten zu kaufen. -- 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: How does swap space work ?
Bernard Wu wrote: > Thats exactly my point. The high priority swap target is not full, > not even half full, yet the lower priority swap file gets used. It was full when the swap slots have been assigned. That does not mean it has to be full still at the time you looked at it. -- Carsten Otte has stopped smoking: Ich habe in 9 Monate, 1 Woche und 4 Tage schon 1.381,39 Euro gespart anstatt 5.755,82 Zigaretten zu kaufen -- 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: How does swap space work ?
Bernard Wu wrote: We are running zVM 5.2 and SLES9-SP3. We have 2 swap files defined to each guest. 1 is VDISK and the other is to minidisk. The VDISK has priority 99 and the dasd has a priority of 42. I've noticed that the 2nd swap space gets used even though the utilization of the VDISK swap space does not reach their allocated size: FilenameTypeSize UsedPriority /dev/dasdb1 partition 259956 111676 99 /dev/dasdc1 partition 215896 281242 I was were under the impression that the second swap file will only get used when the primary swap file is completely filled. Guess I was wrong. Can anyone enlighten me on how it is supposed to work ? The high prio swap target gets filled first. Only when that one is full, the second swap target gets filled. However, if the fist swap target gets free space again existing swap slots that live on the low prio target do not get moved back on the high prio target. When a page is swapped out for the very first time, it gets a swap slot assigned that it will be assigned to over its entire lifetime - even if it gets paged into memory and paged out again. so long. Carsten -- Carsten Otte has stopped smoking: Ich habe in 9 Monate, 1 Woche und 3 Tage schon 1.377,75 Euro gespart anstatt 5.740,66 Zigaretten zu kaufen. -- 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: what is the conversion from mips to Ghz or back
Martin Schwidefsky wrote: > Makes sense, doesn't it ? To me, it does. Interresting default value :-). -- Carsten Otte has stopped smoking: Ich habe in 5 Monate, 2 Tage und 23 Stunden schon 748,66 Euro gespart anstatt 3.119,44 Zigaretten zu kaufen. -- 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: Backing up zLinux
Mark Post wrote: I think you just got yourself into trouble here. I would hardly characterize z/OS as having a "primitive" I/O stack or architecture. Lots of buffering and caching go on there, both in hardware and software. The _real_ difference is that z/OS, just like Linux or z/VM, _always_ has a consistent view of its own data. In a shared DASD environment, this is enforced via serialization techniques, either hardware reserve/release, or software such as GRS, MIM, etc. Without those, backing up one z/OS system from another one would run into similar (but perhaps not as severe) problems with inconsistent data winding up on tape. Yea I guess so, but not by intention. The wording was not meant to imply "simple" or "bad". I just intended to state, that due to caching the consistent view of the data that Linux has is not permanently reflected on disk. And that is causes trouble when backing up from outside that guest. Carsten -- 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: what is the conversion from mips to Ghz or back
Wouldn't be the Linux bogomips a good comparison parameter, since it comes in all Linux flavors? As the name indicates, the "bogus" bogomips rating is a very good indicator for comparing CPU performance. See this HowTo document for details: http://tldp.org/HOWTO/html_single/BogoMips/ so long, Carsten -- 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: Backing up zLinux
David Boyes wrote: Other applications, such as FTP server, NSF server, Samba server, print servers, I wouldn't think would care. You would lose what was put on them during the backup process (hence making the backups as short as possible is a good thing). See above. I'd rather engineer a solution that doesn't depend on exceptions. I would like to second David's point here. The whole story does not depend on what running is, and neither does it depend on the application. As long as you have a file system mounted read+write, you get inconsistent data when taking a snapshot from *outside* Linux. I know this is díferent with operating systems that have a more primitive IO stack like z/OS, which don't do caching and write behind. For Linux, do always use dm-snapshot or a backup client _inside_ the machine or mount read-only. cheers, Carsten -- Carsten Otte has stopped smoking: Ich habe in 5 Monate, 1 Tag und 18 Stunden schon 742,93 Euro gespart anstatt 3.095,57 Zigaretten zu kaufen. -- 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: BUG: Soft Lockup detected on CPU#
Martin wrote: The way how the soft-lockup detection works right now is broken for system that utilize virtual cpus. You could argue that all zSeries systems use virtualized cpu so the feature does not make sense. With dedicated PUs on a logical partition, and when running on raw iron the feature should work fine afaict. No? cheers, Carsten -- Carsten Otte has stopped smoking: Ich habe in 4 Monate, 2 Wochen und 1 Tag schon 665,94 Euro gespart anstatt 2.774,76 Zigaretten zu kaufen. -- 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: Soft Lockup detected on CPU#
Post, Mark K wrote: > Then I would say you should not be trying to use vanilla kernel > sources from kernel.org. The vanilla tree is pretty up to date these days. It is good for all of us when it gets some testing, before the shiny new features hit our production systems via a new distribution. Fortunately the new 2.6.x. kernel development process, and Martins personal commitment, gets us rid of the need to apply IBMs patch sets to make things work proper. Vanillas is also a nice vehicle that allows users to preview what to expect in the next SuSE/RedHat release. Carsten -- Carsten Otte has stopped smoking: Ich habe in 4 Monate, 2 Wochen und 17 Stunden schon 661,12 Euro gespart anstatt 2.754,69 Zigaretten zu kaufen. -- 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: zLinux User Passwords on console
Alan Altmark wrote: > Today we tout > the mainframe as being best-of-breed for context switching. But > what triggers those context switches? Interrupts, of course. > Timers. I believe you are referring to Linux context switches, as measured with lmbench. The mainframe is very fast with those indeed, but the benchmark is about switching from one user context to another user context, which is done once scheduler decides to run a different task. In Linux, interrupts are not called context switch, just "interrupt". Btw: some interrupts lead to a context switch, when a user process was waiting for an event (disk read via read() syscall for example). The interrupt handler wakes the user context, and the scheduler is likely to run this task for fairness reason: the task has not computed for some time, it was waiting for I/O to complete while other tasks enjoyed the CPU horsepower. cheers, Carsten -- Carsten Otte has stopped smoking: Ich habe in 4 Monate, 1 Woche und 2 Tage schon 636,45 Euro gespart anstatt 2.651,88 Zigaretten zu kaufen. -- 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: SLES9 out of support?
Jim Elliott <[EMAIL PROTECTED]> wrote: > IBM no longer provides updates for Linux for 31-bit (for the 2.6.16 and later kernels). Carsten Otte wrote: This statement clearly is not true. In fact, we do. Here is an example to prove my statement. Note the diff from Martin against head31.S which is the kernel startup code used for 31-bit kernel only. This code is heading for integration into 2.6.19+, and directed yield is a feature new in future Linux distributions. http://marc.theaimsgroup.com/?l=linux-arch&m=115858205621775&w=2 cheers, Carsten -- Carsten Otte has stopped smoking: Ich habe in 3 Monate, 3 Wochen und 4 Tage schon 564,91 Euro gespart anstatt 2.353,80 Zigaretten zu kaufen. -- 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: SLES9 out of support?
Jim Elliott <[EMAIL PROTECTED]> wrote: IBM no longer provides updates for Linux for 31-bit (for the 2.6.16 and later kernels). This statement clearly is not true. In fact, we do. Carsten -- Carsten Otte has stopped smoking: Ich habe in 3 Monate, 3 Wochen und 4 Tage schon 564,87 Euro gespart anstatt 2.353,65 Zigaretten zu kaufen. -- 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: collaborative memory management on Linux weekly news
Carsten Otte wrote: This week's Linux weekly news features an article about the second stage of our collaborative memory management technology with z/VM. Subscribers can read the article now, it will become public content next Wednesday: http://lwn.net/Articles/198380/ The article is now public. -- Carsten Otte has stopped smoking: Ich habe in 3 Monate, 2 Wochen und 6 Tage schon 541,11 Euro gespart anstatt 2.254,62 Zigaretten zu kaufen. -- 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: mknod sles8 - got to 256
Marcy Cortes wrote: I got as far as mknod -m 600 /dev/dasdbl3 b 94 255 I need to add some more dasd. Do I just use 95? If in doubt, see /proc/partitions. The dasd driver allocates spare major numbers. have fun, Carsten -- Carsten Otte has stopped smoking: Ich habe in 3 Monate, 2 Wochen und 6 Tage schon 541,09 Euro gespart anstatt 2.254,58 Zigaretten zu kaufen. -- 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: swap partition increase
LJ Mace wrote: I know the next statement isn't a permanent fix but I need to know if(and how) I can increase my swap space. You can prepare another dasd with dasdfmt and fdasd, then use mkswap to create a swap space on a partition you created, and activate it using swapon. All commands have man-pages that explain their proper usage. If you want to use the swap space on the next startup, add the disk to /etc/fstab just like the swap partition you do already use (see "man fstab"). cheers, Carsten -- Carsten Otte has stopped smoking: Ich habe in 3 Monate, 2 Wochen und 4 Tage schon 531,81 Euro gespart anstatt 2.215,90 Zigaretten zu kaufen. -- 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: FCP over ECKD performance advantage - why?
Pieter Harder wrote: there are a number of sources that indicate that FCP attached DASD performs better than classic ECKD DASD. My own numbers seem to confirm this. But I am wondering what exactly is the advantage that FCP has over ECKD? The major difference can be found on the protocol level. FCP is a queueing protocol that allows to submit more I/O while the device is processing. The same effect can be archived by doing PAV with FICON devices. Carsten thinks, that both protocols are equally fast enough for not being a bottleneck when set up proper. cheers, Carsten -- Carsten Otte has stopped smoking: Ich habe in 3 Monate, 2 Wochen und 3 Tage schon 527,27 Euro gespart anstatt 2.196,97 Zigaretten zu kaufen. -- 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: How to signal a Linux guest from z/VM?
Thomas Kern wrote: Is hcp/vmcp anymore sensitive in a class G (or less) linux service virtual machine than 'shutdown -h now'? Does anyone really let untrusted users have root access in production service virtual machines? Untrusted users may still issue cp commands, if the admin of the guest machine lets them do that. sudo is the tool of choice to allow user "tapeman" to attach and detach his tape drive while keeping him from all the other stuff one can do with the cp command interface. cheers, Carsten -- Carsten Otte has stopped smoking: Ich habe in 3 Monate, 1 Woche und 6 Tage schon 508,55 Euro gespart anstatt 2.118,96 Zigaretten zu kaufen. -- 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: SLE8 and z/9
John Summerfied wrote: Until answer is satisfactory Ask 'em, "Why?" Because it has not been tested on it. On the other hand, Carsten thinks it works fine. cheers, Carsten -- Carsten Otte has stopped smoking: Ich habe in 3 Monate, 1 Woche und 6 Tage schon 507,38 Euro gespart anstatt 2.114,08 Zigaretten zu kaufen. -- 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
collaborative memory management on Linux weekly news
This week's Linux weekly news features an article about the second stage of our collaborative memory management technology with z/VM. Subscribers can read the article now, it will become public content next Wednesday: http://lwn.net/Articles/198380/ -- Carsten Otte has stopped smoking: Ich habe in 3 Monate, 1 Woche und 6 Tage schon 507,35 Euro gespart anstatt 2.113,96 Zigaretten zu kaufen. -- 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: How to signal a Linux guest from z/VM?
Rich Smrcina wrote: Is that why tape devices aren't enabled when they are attached? Yes. -- Carsten Otte has stopped smoking: Ich habe in 3 Monate, 1 Woche und 6 Tage schon 504,10 Euro gespart anstatt 2.100,42 Zigaretten zu kaufen -- 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: Swap partition "filling up" on RHEL4
Hall, Ken (GTI) wrote: Here's how it looks NOW. Can't speak for the time of the failure. You stated, one swap device is full and another is not (at the time of failure). Where do you get that data from? cheers, Carsten -- Carsten Otte has stopped smoking: Ich habe in 3 Monate, 5 Tage und 22 Stunden schon 470,07 Euro gespart anstatt 1.958,63 Zigaretten zu kaufen. -- 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: Swap partition "filling up" on RHEL4
Carsten Otte wrote: Malloc fails when there is no more free memory except the emergency pool and no swap slots are free. Note that anonymous pages in memory may well occupy both a swap slot and a pysical page in swap cache at the same time. May we see /proc/meminfo content of the system? Carsten -- 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: Swap partition "filling up" on RHEL4
Hall, Ken (GTI) wrote: 1) why we suddenly should have run out of virtual memory (if we did, because we never even got close to that before), and Most probably because your application has sudden demand for anonymous memory. Your report indicates the processes are hanging in "malloc". Maybe an infinite loop containing a memory leak? 2) why it was reported in this way. Does anyone know if there's anything in the swap space management mechanism that would cause Linux to fail malloc if a single swap partition fills up? Malloc fails when there is no more free memory except the emergency pool and no swap slots are free. Note that anonymous pages in memory may well occupy both a swap slot and a pysical page in swap cache at the same time. Is there some kind of process affinity for paging to swap spaces? No. If malloc fails, all swap slots are in use. cheers, Carsten -- 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: Filesystem type and re-sizing filesystems
Rob van der Heij wrote: I hope this does not make you start smoking again, but: lrobv1:~ # resize2fs /dev/dasdd1 resize2fs 1.36 (05-Feb-2005) /dev/dasdd1 is mounted; can't resize a mounted filesystem! lrobv1:~ # cat /proc/mounts | grep dasdd1 /dev/dasdd1 /mnt/0153 ext2 rw,nogrpid 0 0 see http://lwn.net/Articles/89560/ -- Carsten Otte has stopped smoking: Ich habe in 3 Monate, 4 Tage und 32 Minuten schon 460,90 Euro gespart anstatt 1.920,44 Zigaretten zu kaufen. -- 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: Filesystem type and re-sizing filesystems
Carsten Otte wrote: I don't know what you mean by "dynamic", but you can resize ext3 filesystems with resize2fs. There is also a patch that allows online resizing of ext3 filesystems (that is, resizing while the filesystem is mounted). I don't know about the state of it, therefore I would avoid to use it in production environments for the time being. Resize2fs should be part of the e2fsprogs package in sles as far as I know (I don't have a Sles at hand to verify that). To avoid misinterpretion: I would feel safe to use resize2fs after backup of a production system, then do an e2fsck -f before going production again. Online resize of ext3 feels too hot for me to use in production systems, but that does also apply to reiserfs in general. with kind regards, Carsten -- Carsten Otte has stopped smoking: Ich habe in 3 Monate, 4 Tage und 10 Minuten schon 460,83 Euro gespart anstatt 1.920,15 Zigaretten zu kaufen. -- 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: Filesystem type and re-sizing filesystems
Bernard Wu wrote: > We are currently running SLES9-SP3. All our filesystems are type "EXT3" . > Aside from "REISERFS" , are there other filesystem types that will allow > for dynamic re-sizing ? I don't know what you mean by "dynamic", but you can resize ext3 filesystems with resize2fs. There is also a patch that allows online resizing of ext3 filesystems (that is, resizing while the filesystem is mounted). I don't know about the state of it, therefore I would avoid to use it in production environments for the time being. Resize2fs should be part of the e2fsprogs package in sles as far as I know (I don't have a Sles at hand to verify that). with kind regards, Carsten -- Carsten Otte has stopped smoking: Ich habe in 3 Monate, 4 Tage und 58 Sekunden schon 460,80 Euro gespart anstatt 1.920,01 Zigaretten zu kaufen. -- 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: /dev/random
Arty Ecock wrote: >I tried to use gpg today to generate a pgp key on SLES9x. The process > hangs while reading from /dev/random (which is lacking entropy). Hasn't > this whole /dev/random (and /dev/urandom) thing been beaten to death? Does > anyone else use gpg on s390x? Our major source of entropy is the dasd, we don't have a mouse you can move around or a keyboard attached to the "console" on s390. Do some I/O to get more entropy ;-). Carsten -- Carsten Otte has stopped smoking 2 months, 4 weeks and 15 hours ago. He would have spent 430.31 Euro to buy 1,792.98 cigarettes by now. -- 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: XFS and LVM
Yu Safin wrote: > However, XFS is used effectively in aix. Yes, but there XFS means X-Font-Server. While Linux also uses an X-Font-Server, the ancronym decodes to a file system ported from SGI's Irix operating system by Christoph Hellwig and others. -- Carsten Otte has stopped smoking 2 months, 1 week and 1 day ago. He would have spent 335.58 Euro to buy 1,398.28 cigarettes by now. -- 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: SLES10 Abend During Install
Post, Mark K wrote: > illegal operation: 0001 [#1] > CPU:0Not tainted > Process init (pid: 1, task: 007c1748, ksp: 007c3ca0) > Krnl PSW : 070420018000 0002 (0x2) > Krnl GPRS: 0001 03cdb2a8 > 0085 >007c3cf0 0001 010c6160 > 0001 >03cdb3c8 000a3898 03cdb3c8 > >0085 00446888 0025970c > 007c3b38 > Krnl Code: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 > Call Trace: > ([<00259434>] mpage_writepages+0x23c/0xb40) > [<001b07ca>] do_writepages+0x5a/0x74 > [<00256a8c>] sync_inodes+0x540/0x1204 > [<001fdad4>] sys_sync+0x30/0x84 > [<0010f774>] sysc_noemu+0x10/0x16 > [<021464aa>] 0x21464aa > > <0>Kernel panic - not syncing: Attempted to kill init! > HCPGIR450W CP entered; disabled wait PSW 00020001 8000 > 0014259E *Ouch*. In mpage_writepages()? This is pretty much in the very core kernel near I/O page cache and file system, not somewhere near networking. Both CTC device driver and all user space components like Yast avert suspicion as far as I can tell. Is there another message right before "illegal operation: 0001 [#1]"? Probably one that says "kernel BUG" or "Ooops" or similar? Usually kernel traps use illegal opcodes to crash the thing on purpose, but they are supposed to print a useful message before that. cheers, Carsten -- 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: Bad Linux backups
J Leslie Turriff wrote: > Okay, now, wait; are you saying that the storage device _does_ have a > mechanism for communicating with the Linux filesystem to determine what > filesystem pages are still cached in main storage and have not yet been > commited to external storage? No, it does not. Invention required. cheers, Carsten -- 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: Bad Linux backups
> On Wednesday, 07/26/2006 at 10:33 EST, J Leslie Turriff > <[EMAIL PROTECTED]> wrote: >> Sounds to me, then, like the use of the >> snapshot/mirror/peer-to-peer copy features of storage devices e.g. >> Shark, SATABeast, etc. are currently dangerous to use with Linux >> filesystems. They would need to be able to coordinate their activities >> with the filesystem lock/unlock components of the kernel to be made >> safe? Alan Altmark wrote: > No, they are not "currently dangerous to use with Linux". The > snapshot/flashcopy features provide a point-in-time consistent view of an > entire device or range of blocks/cylinders. In a "normal" track-by-track > read, data on the device can change while you're reading. I am sorry, but I have to disagree with Alan's statement. They _are_ currently dangerous to use with Linux volumes that are being accessed _because_ unlike dm-snapshot the filesystem is not frozen in Linux (lockfs) and thus the data on disk is inconsitent due to caching. DM-snapshot does the desired trick, flashcopy does not. I feel sorry for causing confusion by creating the expectation that flashcopy can be used to snapshot linux volumes before. cheers, Carsten -- 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: Bad Linux backups
J Leslie Turriff wrote: > Sounds to me, then, like the use of the > snapshot/mirror/peer-to-peer copy features of storage devices e.g. > Shark, SATABeast, etc. are currently dangerous to use with Linux > filesystems. They would need to be able to coordinate their activities > with the filesystem lock/unlock components of the kernel to be made > safe? Exactly, yes. cheers, Carsten -- 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: Bad Linux backups
Christoph Hellwig wrote: > But that's not how snapshot work. When you do a snapshot the filesystem > is frozen. That means: new file writers are blocked from dirtying the > filesystem throug the pagecache. The filesystem block callers that want > to create new transactions. Then the whole file cache is written out > and the asynchronous write ahead log (journal) is written out on disk. > The filesystem is in a fully consistant state. Trust me, I've > implemented this myself for XFS. Very interresting indeed. This pointed me to reading the lockfs/unlockfs semantics in Linux, and I think I need to withdraw my statement regarding flashcopy snapshots: because of the fact that there is no lockfs/unlockfs interaction when doing flashcopy, and because of dirty pages in the page cache during snapshot, flashcopy will not generate a consistent snapshot. Therefore, using flashcopy on an active volume from outside Linux is _not_ suitable for backup purposes. The only feasible way to get a consistent snapshot is to use dm-snapshot from within Linux. This snapshot copy can later on be used with a backup feature outside Linux. regards, Carsten -- 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