Re: XIV Questions

[Raymond Higgs]

John,

I think part of your problem is the set edev command.  You should be using 
a wwpn of your XIV, not the channel.

Regards,

Ray Higgs
System z FCP Firmware Development
Bld. 706, B42
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
rayhi...@us.ibm.com

The IBM z/VM Operating System  wrote on 
06/23/2011 03:25:45 PM:

> From: John Hanley 
> To: IBMVM@LISTSERV.UARK.EDU
> Date: 06/23/2011 03:27 PM
> Subject: Re: XIV Questions
> Sent by: The IBM z/VM Operating System 
> 
> Hi Ray,
> Yes I'm using NPIV and it looks on the z side.
> 
> q fcp wwpn 2201
> FCP 2201 NPIV WWPN C05076FD1C00021C 
> CHPID E0 PERM WWPN 5005076401226A45 
> FREE 
> 
> set edev c201 type fba attr xiv fcp_dev 2201 wwpn C05076FD1C00021C 
> lun 0002 
> EDEV C201 was created. 
> 
> vary on c201 
> HCPSZP8701I Path FCP_DEV 2201 WWPN C05076FD1C00021C LUN 
> 0002 was del
> eted from EDEV C201 because it is invalid. 
> HCPCPN8700I Emulated Device C201 cannot be varied online because 
> HCPCPN8700I there are no valid paths defined to the device. 
> 1 device(s) specified; 0 device(s) successfully varied online 
> 
> 
> Yet I can attach the rdev 2201 to a linux guest and do a chccwdev -e
> 0.0.220, and it comes online and shows up in the cisco switch with 
> the correct npiv?
> 
> But.even then the XIV does not see it and it is in the zone?
> 
> I think I'm missing very basic.
> 
> I have TS3500 scsi tapes working from this z10 thru the same san 
> just fine, also npiv.
> 
> 
> John Hanley
> 
> [image removed] Raymond Higgs ---06/23/2011 03:04:00 PM---Raymond 
> Higgs 

> 
> Raymond Higgs  
> Sent by: The IBM z/VM Operating System  
> 06/23/2011 03:03 PM 
> 
> Please respond to
> The IBM z/VM Operating System 
> 
> [image removed] 
> To
> 
> [image removed] 
> IBMVM@LISTSERV.UARK.EDU
> 
> [image removed] 
> cc
> 
> [image removed] 
> 
> [image removed] 
> Subject
> 
> [image removed] 
> Re: XIV Questions
> 
> [image removed] 
> 
> [image removed] 
> 
> 
> John, 
> 
> Ficon Express 4 will work. Are you using NPIV? The channel will not 
> login NPIV subchannels to the fabric until the EDEV is varied online. 
> 
> Regards, 
> 
> Ray Higgs
> System z FCP Firmware Development
> Bld. 706, B42
> 2455 South Road
> Poughkeepsie, NY 12601
> (845) 435-8666, T/L 295-8666
> rayhi...@us.ibm.com 
> 
> The IBM z/VM Operating System  wrote on 06/
> 23/2011 12:47:12 PM:
> 
> > From: John Hanley  
> > To: IBMVM@LISTSERV.UARK.EDU 
> > Date: 06/23/2011 12:55 PM 
> > Subject: XIV Questions 
> > Sent by: The IBM z/VM Operating System  
> > 
> > I have a couple of questions on setting up an XIV box to be used by 
our z10.
> > It will be for zLinux under zVM (6.1), and I'd like to configure the
> > LUNs as EDEVs. 
> > Does the XIV require FICON Express8 cards on the z10, or will Express4 
work?
> > The XIV does not see the z10 through the switch. 
> > Does the san switch need to be an IBM "SVC" for EDEVICES with the XIV? 

> > Thanks.

Re: XIV Questions

[Raymond Higgs]

John,

Ficon Express 4 will work.  Are you using NPIV?  The channel will not 
login NPIV subchannels to the fabric until the EDEV is varied online.

Regards,

Ray Higgs
System z FCP Firmware Development
Bld. 706, B42
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
rayhi...@us.ibm.com

The IBM z/VM Operating System  wrote on 
06/23/2011 12:47:12 PM:

> From: John Hanley 
> To: IBMVM@LISTSERV.UARK.EDU
> Date: 06/23/2011 12:55 PM
> Subject: XIV Questions
> Sent by: The IBM z/VM Operating System 
> 
> I have a couple of questions on setting up an XIV box to be used by our 
z10.
> It will be for zLinux under zVM (6.1), and I'd like to configure the
> LUNs as EDEVs. 
> Does the XIV require FICON Express8 cards on the z10, or will Express4 
work?
> The XIV does not see the z10 through the switch. 
> Does the san switch need to be an IBM "SVC" for EDEVICES with the XIV?
> Thanks.

Re: Using EMC Clariion SCSI disk with Linux

[Raymond Higgs]

Anne,

Mixing long and short wave is definitely your problem.  It makes sense 
that the switch isn't negotiating at all.  When the optics/cables are 
mismatched, the ports can't get word sync and can't even start sending 
ordered sets to do speed negotiation.

A quick google search says that long wave is supported by that switch:

http://www.brocade.com/downloads/documents/data_sheets/product_data_sheets/5100-DS-04.pdf

Media types 8 Gbps: Requires Brocade hot-pluggable SFP+, LC connector; 8 
Gbps Short-Wavelength Laser (SWL); 8 Gbps Long-Wavelength Laser (LWL); 
distance depends on fiber-optic cable and port speed

4 Gbps: Requires Brocade hot-pluggable, Small Form-factor Pluggable (SFP), 
LC connector; 4 Gbps Short-Wavelength Laser (SWL); 4 Gbps Long-Wavelength 
Laser (LWL); 4 Gbps Extended Long-Wavelength Laser (ELWL); distance 
depends on fiber-optic cable and port speed

I don't know why he would say that long wave doesn't exist.  Maybe he 
really means that they don't have any long wave in that switch.  Long wave 
optics are usually more expensive...

Our Ficon Express 8 cards support both long and short wave.  Maybe you 
have some other channels that are short wave that you could use instead.

Regards,

Ray Higgs
System z FCP Firmware Development
Bld. 706, B42
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
rayhi...@us.ibm.com

The IBM z/VM Operating System  wrote on 
05/09/2011 11:20:16 AM:

> From: "Crabtree, Anne D" 
> To: IBMVM@LISTSERV.UARK.EDU
> Date: 05/09/2011 11:21 AM
> Subject: Re: Using EMC Clariion SCSI disk with Linux
> Sent by: The IBM z/VM Operating System 
> 
> Accept-Language: en-US
> Content-Language: en-US
> X-MS-Has-Attach:
> X-MS-TNEF-Correlator:
> acceptlanguage: en-US
> Content-Type: multipart/alternative;
> boundary="_000_897C82FC69765D45A301AF8F5D1210CB09BF7F6COTB6MAIL01execu_"
> MIME-Version: 1.0
> X-Sophos-ESA: [mx6.uark.edu] 3.6.9.0, Antispam-Engine: 2.7.2.
> 1390750, Antispam-Data: 2011.5.9.145714
> 
> --_000_897C82FC69765D45A301AF8F5D1210CB09BF7F6COTB6MAIL01execu_
> Content-Type: text/plain; charset="us-ascii"
> Content-Transfer-Encoding: quoted-printable
> 
> 1.The cabling appears fine.
> 
> 2. Our Ficon Express 8 is LX (long wave) and open system guy 
says s=
> witch is:
> 
> Make:  EMC (Brocade)
> 
> Model:  DS-5100
> 
> Firmware:  6.1.0c
> 
> NPIV Enabled Port
> 
> He says optics are short wave and there is no such thing as long wave?
> 
> 3.  He says there are not errors on switch that the port is not 
"negoti=
> ating" at all.
> 
> 4.  As he says above NPIV is enabled on switch.  Fabric Login Status 
ha=
> s *** in all fields.
> 
> 
> So with that being said, it appears the optics are the problem?  Do you 
agr=
> ee that there is no such thing as long wave on open system switches?
> 
> 
> 
> Anne D. Crabtree
> System Programmer
> WV Office of Technology Data Center
> 1900 Kanawha Blvd East
> Charleston, WV  25305
> (304)558-5914 ext 58292
> (304)558-1441 fax
> 
> From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On 
Beh=
> alf Of Raymond Higgs
> Sent: Friday, May 06, 2011 9:12 AM
> Subject: Re: Using EMC Clariion SCSI disk with Linux
> 
> The IBM z/VM Operating System  wrote on 
05/06/2011=
>  07:51:57 AM:
> 
> > From: "Crabtree, Anne D" 
> > To: IBMVM@LISTSERV.UARK.EDU
> > Date: 05/06/2011 07:53 AM
> > Subject: Using EMC Clariion SCSI disk with Linux
> > Sent by: The IBM z/VM Operating System 
> >
> > I posted this on IBM-MAIN with not much response so hoping someone
> > on z/VM might be able to help.  Opened a PMR with IBM software but
> > they feel it is a hardware issue.  Waiting for information on switch
> > before contacting IBM hardware folks...
> >
> > We have a z10 in which I have defined 2 FCP chpids on our FICON
> > Express8 cards.  These chpids have NPIV enabled and I have provided
> > the open system team with the WWPN's that were generated.  The FCP
> > chpids are dedicated to our 1 Linux IFL.  We have connected the
> > fiber, but the chpid, while staying ONLINE, gets a SEQUENCE TIME-OUT
> > on the HMC.  The light is on on the mainframe side, but not on the
> > switch.  The open system team tells me that there is nothing for
> > them to do on their side until they can "see" the mainframe (light
> > comes on?).  In z/VM, the chips appear fine, defined as FCP with
> > devices attached and FREE.  I'm not sure what else I can do from the
> > mainframe side to help the situation?  I am not involved in the
> > Linux portion but the Linux guy says he can't "s

Re: Using EMC Clariion SCSI disk with Linux

[Raymond Higgs]

The IBM z/VM Operating System  wrote on 
05/06/2011 12:58:23 PM:

> From: "Crabtree, Anne D" 
> To: IBMVM@LISTSERV.UARK.EDU
> Date: 05/06/2011 01:02 PM
> Subject: Re: Using EMC Clariion SCSI disk with Linux
> Sent by: The IBM z/VM Operating System 
> 
> Thank you Ray.
> Will verify these things and get back to you (probably next week). 
> 
> Thank you Heiko.
> Will research these as well.  Can you tell me where speed is defined?
> 
>  I appreciate the responses. 
> 
> From: The IBM z/VM Operating System [IBMVM@LISTSERV.UARK.EDU] On 
> Behalf Of Raymond Higgs [rayhi...@us.ibm.com]
> Sent: Friday, May 06, 2011 9:11 AM
> Subject: Re: Using EMC Clariion SCSI disk with Linux

> The IBM z/VM Operating System  wrote on 05/
> 06/2011 07:51:57 AM:
> 
> > From: "Crabtree, Anne D"  
> > To: IBMVM@LISTSERV.UARK.EDU 
> > Date: 05/06/2011 07:53 AM 
> > Subject: Using EMC Clariion SCSI disk with Linux 
> > Sent by: The IBM z/VM Operating System  
> > 
> > I posted this on IBM-MAIN with not much response so hoping someone 
> > on z/VM might be able to help.  Opened a PMR with IBM software but 
> > they feel it is a hardware issue.  Waiting for information on switch
> > before contacting IBM hardware folks... 
> > 
> > We have a z10 in which I have defined 2 FCP chpids on our FICON 
> > Express8 cards.  These chpids have NPIV enabled and I have provided 
> > the open system team with the WWPN's that were generated.  The FCP 
> > chpids are dedicated to our 1 Linux IFL.  We have connected the 
> > fiber, but the chpid, while staying ONLINE, gets a SEQUENCE TIME-OUT
> > on the HMC.  The light is on on the mainframe side, but not on the 
> > switch.  The open system team tells me that there is nothing for 
> > them to do on their side until they can "see" the mainframe (light 
> > comes on?).  In z/VM, the chips appear fine, defined as FCP with 
> > devices attached and FREE.  I'm not sure what else I can do from the
> > mainframe side to help the situation?  I am not involved in the 
> > Linux portion but the Linux guy says he can't "see" the luns from 
> > Linux.  Both sides appear to be waiting on me to fix things!
> > 
> > The SCSI is a "CLARiiON  CX-700 with FLARE (operating system) 
> > version 26 revision 31 (02.26.700.5.031) which is the latest release
> > for the CX-700 product line. " 
> 
> Anne, 
> 
> The bad news is that we don't support EMC Clariion, but I have heard
> accounts of it working just fine.  Since you are having problems 
> between the channel and switch, you still will be able to get some 
> IBM support.  I'd check these things: 
> 
> * Double check the cabling between the channel and the switch.  Is 
> the channel really plugged into the switch port that you think it is? 
> * Make sure that the optics in the switch, the optics in the 
> channel, and the fibre are either all long wave or all short wave.  No 
mixing.
> * Check the switch for bit errors.  If the switch is seeing bit 
> errors, it may have disabled the port on you.  Or even simpler, was 
> the switch port ever online? 
> * Is NPIV enabled on your switch?  There is a panel on the SE called
> Fabric Login Status that may have some clues...  Use the HMC to do 
> single object operatings to your z10 CPC.  Navigate to your channel.
> It doesn't matter if it is the chpid or pchid.  Find "channel 
> operations" in the menus (I'm used to classic mode so I use those 
> arrows to cycle through the right hand panel.  Use "channel problem 
> determination".  You might have to select an lpar at this point. 
> Then "Fabric login Status...".  This panel shows a bunch of fibre 
> channel info from the channel about how it is logged into your 
> switch fabric.  There might be some error messages here. 
> 
> Regards, 
> 
> Ray Higgs
> System z FCP Firmware Development
> Bld. 706, B42
> 2455 South Road
> Poughkeepsie, NY 12601
> (845) 435-8666,  T/L 295-8666
> rayhi...@us.ibm.com

Anne,

Your Ficon Express 8 hardware will try to negotiate to 8, 4 or 2 gig. 
There is no way on the SE/HMC to force a channel to certain speed.  All of 
the switch consoles have speed settings in different places...

Regards,

Ray Higgs
System z FCP Firmware Development
Bld. 706, B42
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
rayhi...@us.ibm.com

Re: Using EMC Clariion SCSI disk with Linux

[Raymond Higgs]

The IBM z/VM Operating System  wrote on 
05/06/2011 07:51:57 AM:

> From: "Crabtree, Anne D" 
> To: IBMVM@LISTSERV.UARK.EDU
> Date: 05/06/2011 07:53 AM
> Subject: Using EMC Clariion SCSI disk with Linux
> Sent by: The IBM z/VM Operating System 
> 
> I posted this on IBM-MAIN with not much response so hoping someone 
> on z/VM might be able to help.  Opened a PMR with IBM software but 
> they feel it is a hardware issue.  Waiting for information on switch
> before contacting IBM hardware folks... 
> 
> We have a z10 in which I have defined 2 FCP chpids on our FICON 
> Express8 cards.  These chpids have NPIV enabled and I have provided 
> the open system team with the WWPN's that were generated.  The FCP 
> chpids are dedicated to our 1 Linux IFL.  We have connected the 
> fiber, but the chpid, while staying ONLINE, gets a SEQUENCE TIME-OUT
> on the HMC.  The light is on on the mainframe side, but not on the 
> switch.  The open system team tells me that there is nothing for 
> them to do on their side until they can "see" the mainframe (light 
> comes on?).  In z/VM, the chips appear fine, defined as FCP with 
> devices attached and FREE.  I'm not sure what else I can do from the
> mainframe side to help the situation?  I am not involved in the 
> Linux portion but the Linux guy says he can't "see" the luns from 
> Linux.  Both sides appear to be waiting on me to fix things!
> 
> The SCSI is a "CLARiiON  CX-700 with FLARE (operating system) 
> version 26 revision 31 (02.26.700.5.031) which is the latest release
> for the CX-700 product line. "

Anne,

The bad news is that we don't support EMC Clariion, but I have heard 
accounts of it working just fine.  Since you are having problems between 
the channel and switch, you still will be able to get some IBM support. 
I'd check these things:

* Double check the cabling between the channel and the switch.  Is the 
channel really plugged into the switch port that you think it is?
* Make sure that the optics in the switch, the optics in the channel, and 
the fibre are either all long wave or all short wave.  No mixing.
* Check the switch for bit errors.  If the switch is seeing bit errors, it 
may have disabled the port on you.  Or even simpler, was the switch port 
ever online?
* Is NPIV enabled on your switch?  There is a panel on the SE called 
Fabric Login Status that may have some clues...  Use the HMC to do single 
object operatings to your z10 CPC.  Navigate to your channel.  It doesn't 
matter if it is the chpid or pchid.  Find "channel operations" in the 
menus (I'm used to classic mode so I use those arrows to cycle through the 
right hand panel.  Use "channel problem determination".  You might have to 
select an lpar at this point.  Then "Fabric login Status...".  This panel 
shows a bunch of fibre channel info from the channel about how it is 
logged into your switch fabric.  There might be some error messages here.

Regards,

Ray Higgs
System z FCP Firmware Development
Bld. 706, B42
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
rayhi...@us.ibm.com

Re: Restrictions in FCP point-to-point topology

[Raymond Higgs]

The IBM z/VM Operating System  wrote on 
04/28/2011 12:19:53 PM:

> From: Alan Ackerman 
> To: IBMVM@listserv.uark.edu
> Date: 04/28/2011 12:20 PM
> Subject: Restrictions in FCP point-to-point topology
> Sent by: The IBM z/VM Operating System 
> 
> Restrictions in FCP point-to-point topology
> 
> We are planning to use FCP disk with a z/Linux under z/VM system,
> running under z/VM 5.4.  To reduce cost, we are planning to use the
> point-to-point topology.
> 
> Are there any restriction in using this topology?
> 
> I seem to remember some restrictions mentioned in the presentation given
> at SHARE when the point-to-point support first appeared in z/VM.
> 
> The load (write) volume will be high (the higher the better), but only
> occasionally and for less than a day at a time.  Another reason to use
> point-to-point is that we think it will perform better without the
> switch.
> 
> We want to have a 4TB LVM, spread across 8 LUNs.
> 
> Picture:
> 
> +---++---+
> |  LPAR 1   ||  LPAR 2   |
> |   +---+|   +---+
> |   |  z/VM 1   |-FCP--4TB---FCP-|   |  z/VM 2   |
> |   |   +---+  disk  |   |   +---+
> |   |   |  Linux 1  ||   |   |  Linux 2  |
> +---+---+---++---+---+---+
> 
> A wider picture is
> 
> z/OS<--Hipersocket-->Linux-->FCP<--Hipersocket-->z/OS
> DB2  disk
> 
> I found a presentation "Introduction to SCSI over FCP for Linux on
> System z" at .  It
> mentions (on foil 6) the two topologies, point-to-point and switched
> fabric, but gives no restrictions.
> 
> Thanks!
> 
> Alan Ackerman
> Alan (dot) Ackerman (at) Bank of America (dot) com
> 

Alan,

Here are some things that I can think of at the moment:

* NPIV is not supported point-to-point, which impacts lun masking (host 
connections).
* Which storage array do you plan on using?  Some arrays do not support 
point-to-point at all, or they support the poor man's arbitrated loop 
topology with exactly 2 l_ports.  Our system z channels support the 
official FC-FS point-to-point topology.  We do not support any arbitrated 
loop topologies.
* We have not measured switched perf vs point-to-point perf here in 
Poughkeepsie.  I think the latency of the switch (probably microseconds) 
will be much less than the latency of real disks (milliseconds). 
Especially with lots of writes.  I don't think you'll notice the 
difference, but that is just my guess.

Regards,

Ray Higgs
System z FCP Firmware Development
Bld. 706, B42
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
rayhi...@us.ibm.com

Re: Cannot see SVC disk zLinux

[Raymond Higgs]

The IBM z/VM Operating System  wrote on 
04/13/2011 10:15:44 AM:

> From: Terry Wilson 
> To: IBMVM@LISTSERV.UARK.EDU
> Date: 04/13/2011 10:19 AM
> Subject: Re: Cannot see SVC disk zLinux
> Sent by: The IBM z/VM Operating System 
> 
> Entered it manually and here is the info. 
> command: 
> zlinux1:/sys/bus/ccw/drivers/zfcp/0.0.5401 # lsscsi -g 
> [0:0:1:1074479120]diskIBM  2107900  .280  /dev/sda 
> /dev/sg0  >> 8300 disk works fine 
> [1:0:2:49409]diskIBM  2145   - /dev/
> sg1  >>  new SVC DISK 
> 
> command: 
> zlinux1:/sys/bus/ccw/drivers/zfcp/0.0.5401 # sg_luns /dev/sg1 
> Report Luns command has bad field in cdb 
> 
> When zoning the LUNS the 8300 has zLinux as a valid unittype the SVCdoes 
not.

Terry,

Aren't you referring to host connections, which is sometimes called lun 
masking?  Here we use type = generic when adding hosts on the SVC console.

Since the sg devices get created, I don't think there is a zoning problem. 
 Linux can obviously get to the target wwpn because it figured out that 
there is a 2145 SVC there.

I think there is a problem with the hosts that were specified on the SVC 
console.  I tried to do lsluns with a channel that is unknown to our SVC 
and got the same messages as you:

root@Laplace-4E15:lsluns -c 0.1.7900 -p 0x5005076801104bac
Scanning for LUNs on adapter 0.1.7900
at port 0x5005076801104bac:
Unable to send the REPORT_LUNS command to LUN.

I then configured the well known lun by hand, and also got the same 
messages as you:

root@Laplace-4E15:sg_luns /dev/sg4
Report Luns command has bad field in cdb

So I ran it again with -v

root@Laplace-4E15:sg_luns -v /dev/sg4
report luns cdb: a0 00 00 00 00 00 00 00 20 00 00 00
report luns:  Fixed format, current;  Sense key: Illegal Request
 Additional sense: Logical unit not supported
  Info fld=0x0 [0]
Report Luns command has bad field in cdb

Those are the real error messages that I was hoping for!

Then someone here reminded me of scsi_logging_level.  I guess I rely on 
finisar traces too much to see all of bits crossing the fibre channel 
links when I have problems.  Anyway, "scsi_logging_level -s --all 7" will 
turn on verbose messages in /var/log/messages.  So running lsluns against 
the troublesome SVC port gives this:

Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530031] sg_open: dev=4, 
flags=0x8802
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530046] sg_add_sfp: 
sfp=0xec3ae000
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530050] sg_build_reserve: 
req_size=32768
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530054] sg_build_indirect: 
buff_size=32768, blk_size=32768
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530063] sg_build_indirect: 
k=0, num=32768, ret_sz=32768
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530067] sg_build_indirect: 
k_use_sg=1, rem_sz=0
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530071] sg_add_sfp: 
bufflen=32768, k_use_sg=1
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530125] sg_ioctl: sg4, 
cmd=0x2285
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530129] 
scsi_block_when_processing_errors: rtn: 1
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530135] sg_common_write:  scsi 
opcode=0xa0, cmd_size=12
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530140] sg_start_req: 
dxfer_len=8192
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530149] sg_link_reserve: 
size=8192
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530174] scsi 0:0:11:49409: 
[sg4] Send: 0xfcac6c78
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530182] scsi 0:0:11:49409: 
[sg4] CDB: Report luns: a0 00 00 00 00 00 00 00 20 00 00 00
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530210] buffer = 
0xfd2cd3b8, bufflen = 8192, queuecommand 0x0049948c
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530227] leaving 
scsi_dispatch_cmnd()
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530410] scsi 0:0:11:49409: 
[sg4] Done: 0xfcac6c78 SUCCESS
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530418] scsi 0:0:11:49409: 
[sg4]  Result: hostbyte=DID_OK driverbyte=DRIVER_OK
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530425] scsi 0:0:11:49409: 
[sg4] CDB: Report luns: a0 00 00 00 00 00 00 00 20 00 00 00
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530452] scsi 0:0:11:49409: 
[sg4]  Sense Key : Illegal Request [current]
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530461] Info fld=0x0
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530465] scsi 0:0:11:49409: 
[sg4]  Add. Sense: Logical unit not supported
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530475] scsi 0:0:11:49409: 
[sg4] scsi host busy 1 failed 0
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530482] scsi 0:0:11:49409: 
Notifying upper driver of completion (result 802)
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530487] 16 sectors total, 8192 
bytes done.
Apr 13 23:38:55 Laplace-4E15 kernel: [31575.530497] sg_cmd_done: sg4, 
pack_id=0, res=0x802
Apr 13 23:38:55 Laplace-4E15 kernel: [31575

Re: Cannot see SVC disk zLinux

[Raymond Higgs]

The IBM z/VM Operating System  wrote on 
04/13/2011 01:32:11 AM:

> From: Mark Post 
> To: IBMVM@LISTSERV.UARK.EDU
> Date: 04/13/2011 01:32 AM
> Subject: Re: Cannot see SVC disk zLinux
> Sent by: The IBM z/VM Operating System 
> 
> >>> On 4/12/2011 at 04:14 PM, Terry 
> Wilson wrote: 
> > Hello,
> >   I am having a problem connecting ZFCP through my SVC. When I issue 
the 
> > lsluns command it comes back with Unable to send the REPORT_LUNS 
command 
> > to LUN.  Any ideas?
> > Attached are commands 
> > zlinux1:~ # lszfcp -P(
> > 0.0.5400/0x500507680140a23a rport-0:0-0
> > 0.0.5400/0x50050763070342f3 rport-0:0-1 > 8300 disk
> > 0.0.5400/0x500507680110a23a rport-0:0-2
> > 0.0.5400/0x500507680140a31b rport-0:0-3
> > 0.0.5400/0x500507680110a31b rport-0:0-4 
> > 0.0.5401/0x500507680110a23a rport-1:0-0
> > 0.0.5401/0x500507680140a31b rport-1:0-1
> > 0.0.5401/0x500507680110a31b rport-1:0-2
> > 0.0.5401/0x500507680140a23a rport-1:0-3
> > 0.0.5401/0x50050763070342f3 rport-1:0-4> 8300 disk
> > 
> >  zlinux1:~ # lszfcp -D
> > 0.0.5400/0x50050763070342f3/0x4010400b 0:0:1:1074479120
> 
> The LUN number (1074479120) looks extremely suspicious.  In 
> hexadecimal, it's only 4 characters, so the maximum value would be 
> 64K. 

1074479120 = 0x400B4010

It's backward?  It might be normal.  I don't look at these Linux scsi 
device ids enough to know.

Ray Higgs
System z FCP Firmware Development
Bld. 706, B42
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
rayhi...@us.ibm.com

Re: Cannot see SVC disk zLinux

[Raymond Higgs]

Terry,

/var/log/messages looks ok.  49409 (0xc101) is the well known LUN to issue 
report_luns against.  So lsluns is doing the extra step to create an sg 
device for these target ports without any storage configured.

As I alluded to in my previous note, I'd try to do this manually.  Either 
by editing a copy of lsluns or by doing it completely by hand.  Maybe like 
this:

lsscsi -g   <>
chccwdev --online  0.0.5401
echo 0x500507680110a31b > /sys/bus/ccw/drivers/zfcp/0.0.5401/port_add 
<>
echo 0x > 
/sys/bus/ccw/drivers/zfcp/0.0.5401/0x500507680110a31b/unit_add
echo 0xc101 > 
/sys/bus/ccw/drivers/zfcp/0.0.5401/0x500507680110a31b/unit_add
lsscsi -g <>
sg_luns /dev/sgXXX   <>

Hopefully, the last step gives you more meaningful error messages.

I wonder who owns this tool now...

Ray Higgs
System z FCP Firmware Development
Bld. 706, B42
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
rayhi...@us.ibm.com

The IBM z/VM Operating System  wrote on 
04/12/2011 05:55:53 PM:

> From: Terry Wilson 
> To: IBMVM@LISTSERV.UARK.EDU
> Date: 04/12/2011 05:56 PM
> Subject: Re: Cannot see SVC disk zLinux
> Sent by: The IBM z/VM Operating System 
> 
> Running Z/VM 6.1 SLES 11 SP1 
> Here is the var/log/messages from the following command 
> zlinux1:/dev # lsluns 
> Scanning for LUNs on adapter 0.0.5400 
> at port 0x50050763070342f3:>8300 

> 0x4010400b 
> at port 0x500507680110a23a: 
> Unable to send the REPORT_LUNS command to LUN. 
> at port 0x500507680110a31b: 
> Unable to send the REPORT_LUNS command to LUN. 
> at port 0x500507680140a23a: 
> Unable to send the REPORT_LUNS command to LUN. 
> at port 0x500507680140a31b: 
> Unable to send the REPORT_LUNS command to LUN. 
> Scanning for LUNs on adapter 0.0.5401 
> at port 0x50050763070342f3: 
> 0x4010400b>8300 
> at port 0x500507680110a23a: 
> Unable to send the REPORT_LUNS command to LUN. 
> at port 0x500507680110a31b: 
> Unable to send the REPORT_LUNS command to LUN. 
> at port 0x500507680140a23a: 
> Unable to send the REPORT_LUNS command to LUN. 
> at port 0x500507680140a31b: 
> Unable to send the REPORT_LUNS command to LUN. 
> var/log/messages 
> Apr 12 15:54:39 zlinux1 kernel: scsi 0:0:2:49409: Direct-Access 
> IBM  2145  PQ: 1 ANSI: 4 
> Apr 12 15:54:39 zlinux1 kernel: scsi 0:0:2:49409: Attached scsi 
> generic sg1 type 0 
> Apr 12 15:54:39 zlinux1 kernel: scsi 0:0:4:49409: Direct-Access 
> IBM  2145  PQ: 1 ANSI: 4 
> Apr 12 15:54:39 zlinux1 kernel: scsi 0:0:4:49409: Attached scsi 
> generic sg1 type 0 
> Apr 12 15:54:39 zlinux1 kernel: scsi 0:0:0:49409: Direct-Access 
> IBM  2145  PQ: 1 ANSI: 4 
> Apr 12 15:54:39 zlinux1 kernel: scsi 0:0:0:49409: Attached scsi 
> generic sg1 type 0 
> Apr 12 15:54:39 zlinux1 kernel: scsi 0:0:3:49409: Direct-Access 
> IBM  2145  PQ: 1 ANSI: 4 
> Apr 12 15:54:39 zlinux1 kernel: scsi 0:0:3:49409: Attached scsi 
> generic sg1 type 0 
> Apr 12 15:54:40 zlinux1 kernel: scsi: unknown device type 30 
> Apr 12 15:54:40 zlinux1 kernel: scsi 1:0:4:49409: Well-known LUN 
> IBM  2107900  .280 PQ: 0 ANSI: 5 
> Apr 12 15:54:40 zlinux1 kernel: scsi 1:0:4:49409: Attached scsi 
> generic sg1 type 30 
> Apr 12 15:54:40 zlinux1 kernel: scsi 1:0:0:49409: Direct-Access 
> IBM  2145  PQ: 1 ANSI: 4 
> Apr 12 15:54:40 zlinux1 kernel: scsi 1:0:0:49409: Attached scsi 
> generic sg1 type 0 
> Apr 12 15:54:40 zlinux1 kernel: scsi 1:0:2:49409: Direct-Access 
> IBM  2145  PQ: 1 ANSI: 4 
> Apr 12 15:54:40 zlinux1 kernel: scsi 1:0:2:49409: Attached scsi 
> generic sg1 type 0 
> Apr 12 15:54:40 zlinux1 kernel: scsi 1:0:3:49409: Direct-Access 
> IBM  2145  PQ: 1 ANSI: 4 
> Apr 12 15:54:40 zlinux1 kernel: scsi 1:0:3:49409: Attached scsi 
> generic sg1 type 0 
> Apr 12 15:54:41 zlinux1 kernel: scsi 1:0:1:49409: Direct-Access 
> IBM  2145  PQ: 1 ANSI: 4 
> Apr 12 15:54:41 zlinux1 kernel: scsi 1:0:1:49409: Attached scsi 
> generic sg1 type 0 
> 
> 
> 
> 
> 
> 
> 
> From:Raymond Higgs  
> To:IBMVM@LISTSERV.UARK.EDU 
> Date:04/12/2011 03:11 PM 
> Subject:Re: Cannot see SVC disk zLinux 
> Sent by:The IBM z/VM Operating System  
> 
> 
> 
> Terry, 
> 
> Anything in /var/log/messages? 
> 
> lsluns is jus

Re: Cannot see SVC disk zLinux

[Raymond Higgs]

Terry,

Anything in /var/log/messages?

lsluns is just a perl script, so I took a peek...  I have lsluns from 
s390tools-1.11.x-1bb17.20110103.s390x so things maybe different in the 
older versions.

Since you don't have a lun configured on the 0.0.5401 0x500507680110a31b 
path, lsluns should try to configure lun zero for you.  Then look up the 
new sg device and run 'sg_luns /dev/sgXXX'.  If you have a line like this 
in lsluns:

 70 my @output = `sg_luns /dev/$sg_dev 2>/dev/null`;

I think if you remove the "2>/dev/null", there will be a much better error 
message.

It seems like my version of lsluns has all of the right error handling, 
but who knows... maybe there is a bug.

Regards,

Ray Higgs
System z FCP Development
Bld. 706, B42
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
rayhi...@us.ibm.com

The IBM z/VM Operating System  wrote on 
04/12/2011 04:14:43 PM:

> From: Terry Wilson 
> To: IBMVM@LISTSERV.UARK.EDU
> Date: 04/12/2011 04:15 PM
> Subject: Cannot see SVC disk zLinux
> Sent by: The IBM z/VM Operating System 
> 
> -0
> 
> Hello,
>   I am having a problem connecting ZFCP through my SVC. When I issue the 

> 
> lsluns command it comes back with Unable to send the REPORT_LUNS command 

> 
> to LUN.  Any ideas?
> Attached are commands 
> zlinux1:~ # lszfcp -P(
> 0.0.5400/0x500507680140a23a rport-0:0-0
> 0.0.5400/0x50050763070342f3 rport-0:0-1 > 8300 disk
> 0.0.5400/0x500507680110a23a rport-0:0-2
> 0.0.5400/0x500507680140a31b rport-0:0-3
> 0.0.5400/0x500507680110a31b rport-0:0-4 
> 0.0.5401/0x500507680110a23a rport-1:0-0
> 0.0.5401/0x500507680140a31b rport-1:0-1
> 0.0.5401/0x500507680110a31b rport-1:0-2
> 0.0.5401/0x500507680140a23a rport-1:0-3
> 0.0.5401/0x50050763070342f3 rport-1:0-4> 8300 disk
> 
>  zlinux1:~ # lszfcp -D
> 0.0.5400/0x50050763070342f3/0x4010400b 0:0:1:1074479120
> zlinux1:~ # lsluns -c 0.0.5401 -p 0x50050763070342f3
> Scanning for LUNs on adapter 0.0.5401
> at port 0x50050763070342f3:
> 0x4010400b
> zlinux1:~ # lsluns -c 0.0.5401 -p 0x500507680110a31b
> Scanning for LUNs on adapter 0.0.5401
> at port 0x500507680110a31b:
> Unable to send the REPORT_LUNS command to LUN.
> zlinux1:~ # lszfcp -D
> 0.0.5400/0x50050763070342f3/0x4010400b 0:0:1:1074479120
> zlinux1:~ # lsscsi
> [0:0:1:1074479120]diskIBM  2107900  .280  /dev/sda

Re: z196 and z/Manager (URM)

[Raymond Higgs]

The IBM z/VM Operating System  wrote on 
04/04/2011 10:18:38 PM:

> From: "Les Geer (607-429-3580)" 
> To: IBMVM@LISTSERV.UARK.EDU
> Date: 04/04/2011 10:19 PM
> Subject: z196 and z/Manager (URM)
> Sent by: The IBM z/VM Operating System 
> 
> >Hi
> >
> >We are planning for a z196 Enterprise Processor. I was wondering 
specifical=
> >ly about z/Manager. I know this is being touted as the Hub of this new 
Ente=
> >rprise Processor.  I also know that a lot of work on behalf of z/VM and 
z/L=
> >inux was done in this area. I understand that z/Manager contains 
ensembles =
> >that allow for z/VM and z/Linux workload management. It is suppose to 
have =
> >the look and feel like WLM for z/OS.  This is something I have hoping 
for. =
> >Hopefully there is meat with the potatoes!
> >
> >What have been some of the experiences with z/Manager to date. Is it 
truly =
> >a Workload Manager in the way of WLM on z/OS? What have been the 
experience=
> >s with working with the HMC to control z/Manager? Is the HMC accessible 
via=
> > the network, VPN etc.?
> >
> >Thanks!
> >
> 
> At a high level, view the presentation at this URL:
> 
> https://w3-03.sso.ibm.com/sales/support/ShowDoc.wss?
> 
docid=H838775I67832D67&node=node=&showDetails=true&sort=date&hitsize=50&ftext=zM
> anager&showDetails=true&sort=date&hitsize=50
> 
> Watch the line split in copy/paste.
> 
> Best Regards,
> Les Geer
> IBM z/VM and Linux Development

Les,

That link is to an intranet site.  It isn't going to be very helpful for 
most people.

Regards,

Ray Higgs
System z FCP Development
Bld. 706, B42
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
rayhi...@us.ibm.com

Re: z10 shark 800 and FCP

[Raymond Higgs]

Carlos,

If you are having problems creating volumes on the ESS800, then this is 
not a problem with your mainframe.  Updating the IOCDS isn't going to help 
create these volumes.

Ray Higgs
System z FCP Development
Bld. 706, B24
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
rayhi...@us.ibm.com

The IBM z/VM Operating System  wrote on 
09/27/2010 12:49:08 PM:

> Carlos Bodra - Pessoal  
> Sent by: The IBM z/VM Operating System 
> 
> 09/27/2010 12:49 PM
> 
> Please respond to
> The IBM z/VM Operating System 
> 
> To
> 
> IBMVM@LISTSERV.UARK.EDU
> 
> cc
> 
> Subject
> 
> Re: z10 shark 800 and FCP
> 
> Ray,
> 
> Sorry for a not so clear definition.
> 
> What happens is that we are trying to define Open System Storage in 
> ESS800 and when we try to allocate dasd space to z10 (open space 
> using FCP defined ports), ESS800 didn´t able to complete
> configuration, so I can´t define a volume inside this space.
> 
> zvm was installed in eckd disks as traditional zvm does, but I´n not
> trying to ipl from fcp dasd yet, since I was not able to define it yet.
> 
> Problem seems to be something missing in definitions or wrong sequence. 
> 
> As I describe in previous email, ports (at ess800 and z10) and fibre
> cable are ok, since if I chance ports to Ficon protocol all works fine. 
> 
> A question arrised now: ports are SX (shot wave) in both sides, 
> cable could be same used for Ficon? Or FC demands a specific cable?

> Carlos Bodra 
> IBM zSeries Certified Specialist 
> Sao Paulo - Brazil
> 
> Em 27/09/2010 12:29, Raymond Higgs escreveu: 
> 
> Carlos, 
> 
> What do you mean by "My problem is before try to bring up zvm 5.4 or
> linux under it."?  I'm confused.  Are you trying to SCSI IPL vm?  If
> you installed a vm system onto ECKD storage on an ESS800, then you 
> have to IPL it with your ficon channels.  If you have a scsi vm 
> system, then use the fcp info that you installed with. 
> 
> Ray Higgs
> System z FCP Development
> Bld. 706, B24
> 2455 South Road
> Poughkeepsie, NY 12601
> (845) 435-8666,  T/L 295-8666
> rayhi...@us.ibm.com 
> 
> The IBM z/VM Operating System  wrote on 09/
> 27/2010 11:12:03 AM:
> 
> > Carlos Bodra - Pessoal  
> > Sent by: The IBM z/VM Operating System 
> > 
> > 09/27/2010 11:12 AM 
> > 
> > Please respond to
> > The IBM z/VM Operating System  
> > 
> > To 
> > 
> > IBMVM@LISTSERV.UARK.EDU 
> > 
> > cc 
> > 
> > Subject 
> > 
> > Re: z10 shark 800 and FCP 
> > 
> > Robert,
> > 
> > Yes, 2109 and 2005 were zonned to allow z10 access. Same zone was 
> > used to test access to intel, so
> > zonned is working ok.
> > 
> > I can´t bring up a linux image, since I´m not able to define storage
> > to be used by this image. My problem is
> > before try to bring up zvm 5.4 or linux under it. 
> > Carlos Bodra 
> > IBM zSeries Certified Specialist 
> > Sao Paulo - Brazil 
> > 
> > Em 27/09/2010 12:00, Robert J Brenneman escreveu: 
> > Is the 2109 zoned to allow the z10 to access the storage device?
> > 
> > Did you IPL a linux guest or LPAR and use it to drive a fabric login
> > or fabric discovery ?
> > 
> > The ESS and DS4700 will only show you the hosts which have contacted
> > them, since they don't do discovery of the host systems. The z10 will
> > not drive any discovery actions in the SAN by itself ( unlike other
> > platforms ) You have to bring up a Linux image and do a SAN fabric /
> > device scan or attempt to configure the FCP storage from z/VM.
> > 

Re: z10 shark 800 and FCP

[Raymond Higgs]

Carlos,

Yes, the short wave optics and fibre that worked as ficon will work as 
FCP.

Ray Higgs
System z FCP Development
Bld. 706, B24
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
rayhi...@us.ibm.com

The IBM z/VM Operating System  wrote on 
09/27/2010 12:49:08 PM:

> Carlos Bodra - Pessoal  
> Sent by: The IBM z/VM Operating System 
> 
> 09/27/2010 12:49 PM
> 
> Please respond to
> The IBM z/VM Operating System 
> 
> To
> 
> IBMVM@LISTSERV.UARK.EDU
> 
> cc
> 
> Subject
> 
> Re: z10 shark 800 and FCP
> 
> Ray,
> 
> Sorry for a not so clear definition.
> 
> What happens is that we are trying to define Open System Storage in 
> ESS800 and when we try to allocate dasd space to z10 (open space 
> using FCP defined ports), ESS800 didn´t able to complete
> configuration, so I can´t define a volume inside this space.
> 
> zvm was installed in eckd disks as traditional zvm does, but I´n not
> trying to ipl from fcp dasd yet, since I was not able to define it yet.
> 
> Problem seems to be something missing in definitions or wrong sequence. 
> 
> As I describe in previous email, ports (at ess800 and z10) and fibre
> cable are ok, since if I chance ports to Ficon protocol all works fine. 
> 
> A question arrised now: ports are SX (shot wave) in both sides, 
> cable could be same used for Ficon? Or FC demands a specific cable?

> Carlos Bodra 
> IBM zSeries Certified Specialist 
> Sao Paulo - Brazil
> 
> Em 27/09/2010 12:29, Raymond Higgs escreveu: 
> 
> Carlos, 
> 
> What do you mean by "My problem is before try to bring up zvm 5.4 or
> linux under it."?  I'm confused.  Are you trying to SCSI IPL vm?  If
> you installed a vm system onto ECKD storage on an ESS800, then you 
> have to IPL it with your ficon channels.  If you have a scsi vm 
> system, then use the fcp info that you installed with. 
> 
> Ray Higgs
> System z FCP Development
> Bld. 706, B24
> 2455 South Road
> Poughkeepsie, NY 12601
> (845) 435-8666,  T/L 295-8666
> rayhi...@us.ibm.com 
> 
> The IBM z/VM Operating System  wrote on 09/
> 27/2010 11:12:03 AM:
> 
> > Carlos Bodra - Pessoal  
> > Sent by: The IBM z/VM Operating System 
> > 
> > 09/27/2010 11:12 AM 
> > 
> > Please respond to
> > The IBM z/VM Operating System  
> > 
> > To 
> > 
> > IBMVM@LISTSERV.UARK.EDU 
> > 
> > cc 
> > 
> > Subject 
> > 
> > Re: z10 shark 800 and FCP 
> > 
> > Robert,
> > 
> > Yes, 2109 and 2005 were zonned to allow z10 access. Same zone was 
> > used to test access to intel, so
> > zonned is working ok.
> > 
> > I can´t bring up a linux image, since I´m not able to define storage
> > to be used by this image. My problem is
> > before try to bring up zvm 5.4 or linux under it. 
> > Carlos Bodra 
> > IBM zSeries Certified Specialist 
> > Sao Paulo - Brazil 
> > 
> > Em 27/09/2010 12:00, Robert J Brenneman escreveu: 
> > Is the 2109 zoned to allow the z10 to access the storage device?
> > 
> > Did you IPL a linux guest or LPAR and use it to drive a fabric login
> > or fabric discovery ?
> > 
> > The ESS and DS4700 will only show you the hosts which have contacted
> > them, since they don't do discovery of the host systems. The z10 will
> > not drive any discovery actions in the SAN by itself ( unlike other
> > platforms ) You have to bring up a Linux image and do a SAN fabric /
> > device scan or attempt to configure the FCP storage from z/VM.
> > 

Re: z10 shark 800 and FCP

[Raymond Higgs]

Carlos,

Yes.

Ray Higgs
System z FCP Development
Bld. 706, B24
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
rayhi...@us.ibm.com

The IBM z/VM Operating System  wrote on 
09/27/2010 12:53:26 PM:

> Carlos Bodra - Pessoal  
> Sent by: The IBM z/VM Operating System 
> 
> 09/27/2010 12:53 PM
> 
> Please respond to
> The IBM z/VM Operating System 
> 
> To
> 
> IBMVM@LISTSERV.UARK.EDU
> 
> cc
> 
> Subject
> 
> Re: z10 shark 800 and FCP
> 
> Ray,
> 
> Just to clarify, can I connect directly z10 port defined as FCP 
> (IOCP) to a ESS800 SW2 defined FCP? (no switch needed)
> Carlos Bodra 
> IBM zSeries Certified Specialist 
> Sao Paulo - Brazil
> 
> Em 27/09/2010 12:20, Raymond Higgs escreveu: 
> 
> Carlos, 
> 
> z10 with ESS800 should work fine point-to-point.  That config was 
> tested a lot here in Poughkeepsie.  We don't have a ds4700.  It 
> should work though. 
> 
> There is some configuration that needs to be done inside linux.  You
> need to tell linux which busid (device number), target wwpn, and lun
> to use.  If you are using Suse, yast has a nice setup wizard.  Have 
> you done any configuration inside of Linux?  The output of lszfcp -
> D, and lsscsi would be helpful. 
> 
> If you are using edevs, then you need to tell vm which device 
> number, target wwpn, and lun to use. 
> 
> Regards, 
> 
> Ray Higgs
> System z FCP Development
> Bld. 706, B24
> 2455 South Road
> Poughkeepsie, NY 12601
> (845) 435-8666,  T/L 295-8666
> rayhi...@us.ibm.com 
> 

> 
> Carlos Bodra - Pessoal  
> Sent by: The IBM z/VM Operating System  
> 09/27/2010 10:53 AM 
> 
> Please respond to
> The IBM z/VM Operating System 
> 
> To
> 
> IBMVM@LISTSERV.UARK.EDU 
> 
> cc
> 
> Subject
> 
> z10 shark 800 and FCP
> 
> 
> 
> 
> Hi
> 
> I´m trying to connect a z10 to a shark using FCP protocol, but 
> without success.
> We did some tests that are describe below:
> 
> 1 - Connect z10 to shark 800 sw2 adapter with both sides defined as 
> Ficon. Works fine
> 
> 2 - Connect z10 to shark 800 sw2 adapter with both sides defined as 
> FCP. Direct connection, no switch envolved. Didn´t work. We aren´t 
> able to define open system storage in ESS Specialist. Seems that 
> shark doesn´t "see" FCP adapter in z10. After reading some material 
> (redbooks) we found information about that a direct connection 
> didn´t work. Is mandatory use a switch.
> 
> 3 - Connect z10 to shark 800 sw2 adapter with both sides defined as 
> FCP and using an old IBM 2109-S16. Didn´t work. We aren´t able to 
> define open system storage in ESS Specialist. Seems that shark 
> doesn´t "see" FCP adapter in z10. 
> 
> 4 - Connect z10 to shark 800 sw2 adapter with both sides defined as 
> FCP and using an IBM 2005-B16. Didn´t work. We aren´t able to define
> open system storage in ESS Specialist. Seems that shark doesn´t 
> "see" FCP adapter in z10. 
> 
> 5 - Connect z10 to DS4700 via FC adapter with z10 define as FCP and 
> storage defined as FC and using an old IBM 2109-S16. Didn´t work. We
> aren´t able to access open dasd. 
> 
> We checked microcode level for shark and seems to be ok. If I 
> connect shark or DS4700 to an intel server, we are able to access 
> open dasd without problems.
> 
> Since z10 with FCP access is new for us, any hint or tip about are 
welcome.
> 
> Thanks in advance.
> -- 
> Carlos Bodra 
> IBM zSeries Certified Specialist 
> Sao Paulo - Brazil 

Re: z10 shark 800 and FCP

[Raymond Higgs]

Carlos,

What do you mean by "My problem is before try to bring up zvm 5.4 or linux 
under it."?  I'm confused.  Are you trying to SCSI IPL vm?  If you 
installed a vm system onto ECKD storage on an ESS800, then you have to IPL 
it with your ficon channels.  If you have a scsi vm system, then use the 
fcp info that you installed with.

Ray Higgs
System z FCP Development
Bld. 706, B24
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
rayhi...@us.ibm.com

The IBM z/VM Operating System  wrote on 
09/27/2010 11:12:03 AM:

> Carlos Bodra - Pessoal  
> Sent by: The IBM z/VM Operating System 
> 
> 09/27/2010 11:12 AM
> 
> Please respond to
> The IBM z/VM Operating System 
> 
> To
> 
> IBMVM@LISTSERV.UARK.EDU
> 
> cc
> 
> Subject
> 
> Re: z10 shark 800 and FCP
> 
> Robert,
> 
> Yes, 2109 and 2005 were zonned to allow z10 access. Same zone was 
> used to test access to intel, so
> zonned is working ok.
> 
> I can´t bring up a linux image, since I´m not able to define storage
> to be used by this image. My problem is
> before try to bring up zvm 5.4 or linux under it.
> Carlos Bodra 
> IBM zSeries Certified Specialist 
> Sao Paulo - Brazil
> 
> Em 27/09/2010 12:00, Robert J Brenneman escreveu: 
> Is the 2109 zoned to allow the z10 to access the storage device?
> 
> Did you IPL a linux guest or LPAR and use it to drive a fabric login
> or fabric discovery ?
> 
> The ESS and DS4700 will only show you the hosts which have contacted
> them, since they don't do discovery of the host systems. The z10 will
> not drive any discovery actions in the SAN by itself ( unlike other
> platforms ) You have to bring up a Linux image and do a SAN fabric /
> device scan or attempt to configure the FCP storage from z/VM.
> 

Re: z10 shark 800 and FCP

[Raymond Higgs]

Carlos,

z10 with ESS800 should work fine point-to-point.  That config was tested a 
lot here in Poughkeepsie.  We don't have a ds4700.  It should work though.

There is some configuration that needs to be done inside linux.  You need 
to tell linux which busid (device number), target wwpn, and lun to use. If 
you are using Suse, yast has a nice setup wizard.  Have you done any 
configuration inside of Linux?  The output of lszfcp -D, and lsscsi would 
be helpful.

If you are using edevs, then you need to tell vm which device number, 
target wwpn, and lun to use.

Regards,

Ray Higgs
System z FCP Development
Bld. 706, B24
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
rayhi...@us.ibm.com



Carlos Bodra - Pessoal  
Sent by: The IBM z/VM Operating System 
09/27/2010 10:53 AM
Please respond to
The IBM z/VM Operating System 


To
IBMVM@LISTSERV.UARK.EDU
cc

Subject
z10 shark 800 and FCP






Hi

I´m trying to connect a z10 to a shark using FCP protocol, but without 
success.
We did some tests that are describe below:

1 - Connect z10 to shark 800 sw2 adapter with both sides defined as Ficon. 
Works fine

2 - Connect z10 to shark 800 sw2 adapter with both sides defined as FCP. 
Direct connection, no switch envolved. Didn´t work. We aren´t able to 
define open system storage in ESS Specialist. Seems that shark doesn´t 
"see" FCP adapter in z10. After reading some material (redbooks) we found 
information about that a direct connection didn´t work. Is mandatory use a 
switch.

3 - Connect z10 to shark 800 sw2 adapter with both sides defined as FCP 
and using an old IBM 2109-S16. Didn´t work. We aren´t able to define open 
system storage in ESS Specialist. Seems that shark doesn´t "see" FCP 
adapter in z10. 

4 - Connect z10 to shark 800 sw2 adapter with both sides defined as FCP 
and using an IBM 2005-B16. Didn´t work. We aren´t able to define open 
system storage in ESS Specialist. Seems that shark doesn´t "see" FCP 
adapter in z10. 

5 - Connect z10 to DS4700 via FC adapter with z10 define as FCP and 
storage defined as FC and using an old IBM 2109-S16. Didn´t work. We 
aren´t able to access open dasd. 

We checked microcode level for shark and seems to be ok. If I connect 
shark or DS4700 to an intel server, we are able to access open dasd 
without problems.

Since z10 with FCP access is new for us, any hint or tip about are 
welcome.

Thanks in advance.
-- 
Carlos Bodra 
IBM zSeries Certified Specialist 
Sao Paulo - Brazil

Re: OpenSystems FBA question

[Raymond Higgs]

The IBM z/VM Operating System  wrote on 
06/03/2010 09:21:52 AM:

> Mark Pace  
> Sent by: The IBM z/VM Operating System 
> 
> 06/03/2010 09:21 AM
> 
> Please respond to
> The IBM z/VM Operating System 
> 
> To
> 
> IBMVM@LISTSERV.UARK.EDU
> 
> cc
> 
> Subject
> 
> OpenSystems FBA question
> 
> I didn't realize until just recently that with an OSA Express you 
> can share the same 3 addresses amongst all LPARs.
> 
> Does the same apply to FBA devices on DS8xxx?  Can I share the 
> addresses between LPARs or does each LPAR need to use unique 
> address?  Currently I have 32 devices defined on an FCP CHPID and I 
> make sure that different addresses are using in each LPAR.  
> 
> -- 
> Mark Pace
> Mainline Information Systems
> 1700 Summit Lake Drive
> Tallahassee, FL. 32317


Yes, device numbers get mapped to unique subchannels.

Ray Higgs
System z FCP Development
Bld. 706, B24
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
rayhi...@us.ibm.com

Re: Mapping an Emulated DASD device to an SVC LUN number

[Raymond Higgs]

The IBM z/VM Operating System  wrote on 
05/25/2010 03:45:56 PM:

> "Wiggins, Mark"  
> Sent by: The IBM z/VM Operating System 
> 
> 05/25/2010 03:45 PM
> 
> Please respond to
> The IBM z/VM Operating System 
> 
> To
> 
> IBMVM@LISTSERV.UARK.EDU
> 
> cc
> 
> Subject
> 
> Mapping an Emulated DASD device to an SVC LUN number
> 
> We?re looking to migrate to a new z/10 and we?ve come up with an 
> issue that hoping someone out there knows how to resolve. A while 
> back when we installed our SVC (2145) we lost the ability to see the
> ?real? LUN number that was being assigned to our z/VM system as an 
> emulated (9336) device. On the SVC side on things you see a 16 
> character LIN id like this; 013D (which we?d only use 
> the last 4 of ?015E?). When you run the SCSIDISC command on z/VM, 
> the LUN that shows up in z/VM in just the next number in sequential 
> order (e.g. if it was the 11 LUN allocated, it would have a LUN id 
> of 000A). Are there any commands from z/VM that can be 
> issued to give you the LUN id as defined within the SVC? A command 
> like ?q edev e011 details? shows
> 
> EDEV E011 TYPE FBA ATTRIBUTES 2145
>   VENDOR:  IBM PRODUCT :  2145 REVISION:   
>   BLOCKSIZE:  512 NUMBER OF BLOCKS:  83886080
>   PATHS:
> FCP_DEV:  A200 WWPN:  50050768011035EE LUN: 
> 000A PREF
>  CONNECTION TYPE:  SWITCHED
> 
> Mark Wiggins
> University of Connecticut
> 860-486-2792


Mark,

I agree with the others.  I think you need to use the SVC's GUI, command 
line interface, or CIM interfaces to get info about managed disks.  The 
CLI is probably the easiest because it only requires an ssh client.  CIM 
probably requires writing some code, and the GUI needs a browser.

Regards,

Ray Higgs
System z FCP Development
Bld. 706, B24
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
rayhi...@us.ibm.com

Re: z/VM not seeing storage from SAN

[Raymond Higgs]

Daniel,

It's much easier to get info about an FCP channel with lszfcp.  You don't 
have to remember those damn sysfs paths that seem to change with each new 
kernel :)

To figure out a subchannel's wwpn do something like this:

r...@m12-205:lszfcp -a -b 0.0.7800
0.0.7800 host64
Bus = "ccw"
availability= "good"
card_version= "0x0006"
cmb_enable  = "0"
cutype  = "1731/03"
devtype = "1732/03"
failed  = "0"
hardware_version= "0x"
in_recovery = "0"
lic_version = "0xbd23"
modalias= "ccw:t1731m03dt1732dm03"
online  = "1"
peer_d_id   = "0x00"
peer_wwnn   = "0x"
peer_wwpn   = "0x"
status  = "0x542a"
uevent  = "DRIVER=zfcp"
Class = "fc_host"
maxframe_size   = "2112 bytes"
node_name   = "0x5005076400c0b742"
permanent_port_name = "0x500507601d80074f" << 
Sent by: The IBM z/VM Operating System 
04/20/2010 02:14 PM
Please respond to
The IBM z/VM Operating System 


To
IBMVM@LISTSERV.UARK.EDU
cc

Subject
Re: z/VM not seeing storage from SAN






No such file exists.. this is the structure of that directory

[root] zvmtestd:/sys/bus/ccw/devices/0.0.5000# ls
availability  devtype   host37   online port_remove  
uevent
card_version  driverin_recovery  peer_d_id  port_rescan
cmb_enablefailedlic_version  peer_wwnn  status
cutypehardware_version  modalias peer_wwpn  subsystem


On Tue, Apr 20, 2010 at 11:24 AM, Richard Troth  
wrote:
I am a little confused about what WWPN goes where in your case. 

Based on prior email, C05076FAE3000400 is the WWPN of your virtual FCP 
adapter (because NPIV is in place).  It starts with a "C".  But 
50060E80042CC20C looks to me more like a storage side WWPN.  What is the 
"real WWPN" of your FCP adapter 5000?  Might start with "5005", though a 
"5006" prefix is certainly also possible.  (Someone who knows FCP better 
than I might know vendor prefixes and things like that.) 

When you had Linux up (instead of CMS), what did you find in these pseudo 
files? 

/sys/bus/ccw/devices/0.0.5000/physical_wwpn 
/sys/bus/ccw/devices/0.0.5000/wwpn 

The latter should be 0xc05076fae3000400.  What is the other? 
(If you did not have NPIV in effect, then the two would match or one be 
zeros.) 

Since you said that you have it working on one or two other systems, how 
to they look by comparison? 

For zoning and masking and all that, it is not clear to me that you need 
to specify the real WWPN (of the HBA, of the FCP adapter) when NPIV is in 
play.  Eric?  What do you say about that?  You certainly DO need to zone 
and mask the storage and the fabric to the virtual WWPN (the one with the 
"C050" prefix). 

-- Rick;   <><






On Tue, Apr 20, 2010 at 09:13, Daniel Tate  wrote:
To answer someone's earlier question, yes, we've seen every redbook 
imaginable related to this, tried to do a direct zfcp attach, set edev.   
I can't give too much info away about our organization but we are using 
cisco san switches, if that makes a difference.  Both the perm and the 
virtual WWPNs are in our zone.  Port security has been toggled off then 
on, the entire CHPID has been toggled as well.   NPIV is on everywhere, 
NPV is off.  Another earlier question was were we doing a lot of 
attaches/detaches, and the answer is yes.  unfortunately its not easy to 
read this debug information - at times i've had to go to the source code 
of the zfcp driver to make an educated guess at what the "perr" and "ferr" 
codes were with codes 0x0d and 0x05 (thanks to the person who clarified 
those).

This is the output from SCSIDISC DEBUG.  We are currently trying to find 
out what THINOP()=-62 and GETLUN()=-112 indicate.

EXCEPTION::Logfile SCSIDISC LOG A already exists. Delete?  (Y/N):
y
INFO::START @ 20 Apr 2010 07:00:58
INFO::No of attached FCP Channels found=1
PLEASE CHOOSE A NO CORRESPONDING TO A FCP DEVICE OR 'QUIT'
 0) All DEVs 1) 5000
0
INFO::Choosen FCP Sub-Channel Device(s)=ALL DEVICES
INFO::FCP SUB-CHANNEL 5000 Initialized
INFO::For Sub Channel 5000 No of WWPNs found=2
FOR FCP SUBCHANNEL 5000
PLEASE CHOOSE A NO CORRESPONDING TO A WWPN OR 'QUIT'
 0) All WWPNs1) C05076FAE3000400
 2) 50060E80042CC20C
0
INFO::For 5000 Choosen WWPN(s)=ALL WWPNs
INFO::FCP SUB-CHANNEL 5000 Re-Initialized
WARNING::WWPN C05076FAE3000400 Could Not be Opened
DEBUG::ERROR RC:C05076FAE3000400 THINOP()=-62
DEBUG::ERROR Open FCP port has failed
WARNING::WWPN C05076FAE3000400 ignored
DEBUG::ERROR RC:C05076FAE3000400 GETLUN()=-112
INFO::FCP SUB-CHANNEL 5000 Re-Initialized
WARNING::WWPN 50060E80042CC20C Could Not be Opened
DEBUG::ERROR RC:50060E80042CC20C THINOP()=-62
DEBUG::ERROR Open FCP port has failed
WARN

Re: z/VM not seeing storage from SAN

[Raymond Higgs]

Daniel,

What happened with your VM problems?

Which config failed?  The one that you got working, or the new one when 
you tried to replicate?

Which version of linux are you using?

>From this linux log, I only see one thing that is interesting "
fsf_prot_status 0x0080".  This is connection initializing. 
There are 3 ways to get this:

1. The fibre channel link is not up.  If you look at the SE, you should 
see online operating when the link is up.  Did someone unplug the fibre 
from the channel?  Are the optics in your channel, the switch, and the 
fibre consistent?  Either all short wave, or all long wave.  Is the switch 
trying to run at a speed that the channel doesn't support?

2. Are you issuing lots of chccwdev --online, chccwdev --offline, chccwdev 
--online,   We'll send this status up if the QDIO and FSF parts of the 
channel get out of sync.  QDIO quickly initializes.  FSF doesn't because 
it has to communicate with the switch, and this status is a way to tell 
the operating system to be patient.

3. Very similar to 2...  We're waiting on FDISC reply from the switch. Are 
there any clues on the fabric login status panel on the SE?  Have you seen 
this redpaper?

http://www.redbooks.ibm.com/redpapers/pdfs/redp4125.pdf

It has some details about the SE panels.

I hope this helps.

Ray Higgs
System z FCP Development
Bld. 706, B24
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
rayhi...@us.ibm.com



Daniel Tate  
Sent by: The IBM z/VM Operating System 
04/19/2010 12:20 PM
Please respond to
The IBM z/VM Operating System 


To
IBMVM@LISTSERV.UARK.EDU
cc

Subject
Re: z/VM not seeing storage from SAN






OK, we *briefly* got the SAN to show a disk, i was able to format, 
everything.. we attempted to replicate with the exact same config, and it 
failed.   Port security has been toggled, WWPNs have been added and 
removed from the zone.  This is becoming incredibly frustrating.   If 
anyone can provide any answers..  The documentation (as far as i have 
found) is so sparse that i had to look at the source code to identify what 
the hex codes mean


This is the output of sys/kernel/debug/s390dbf/zfcp0.0.5000_hba:

timestamp   01271692944:913553125
cpu 00
tag stat
tag2dism
failed  0x00
status_type 0x
status_subtype  0x
queue_designator   
payload   

timestamp   01271692949:933377062
cpu 01
tag resp
tag2perr
fsf_command 0x000d
fsf_reqid   0x1c985000
fsf_seqno   0x
fsf_issued  01271692949:933235988
fsf_prot_status 0x0080
fsf_status  0x
fsf_prot_status_qual     
fsf_status_qual      
fsf_req_status  0x0010
sbal_first  0x00
sbal_last   0x00
sbal_response   0x00
pool0x01
  2807adda 0001 001f 0002 1abea440 
fd0006a0 e000 
  2808bad6 f2318ac4 1abea440 0100 2000 
  
  280b01b1 f2318ace 1abea440 0080 01

timestamp   01271692950:951330062
cpu 00
tag resp
tag2ferr
fsf_command 0x0005


On Fri, Apr 16, 2010 at 5:00 AM, Eric R Farman  wrote:
Daniel, 

SCSIDISC retrieves all registered WWPNs from the switch connected to the 
selected FCP subchannel(s).  If it only reports the one you had attached, 
it doesn't sound like the switch zones are setup properly to let the LPAR 
see the DASD subsystems.  Has the SAN admin registered the physical WWPN 
of the card with the switch zone, or the NPIV WWPNs (displayed on Q FCP 
output)?  The latter is necessary in NPIV environment. 

Regards,
   Eric

Eric Farman
z/VM I/O Development
IBM Endicott, NY



From: 
Daniel Tate  
To: 
IBMVM@LISTSERV.UARK.EDU 
Date: 
04/15/2010 10:50 AM 
Subject: 
Re: z/VM not seeing storage from SAN 
Sent by: 
The IBM z/VM Operating System 




scsidisc only reports those i've attached.  That particular output was 
from a different device (i.e. 5001).. my terminology is a little off. 
Our SAN admin swears it's configured correctly on the switch.



On Thu, Apr 15, 2010 at 9:38 AM, Eric R Farman  wrote: 
Hi Daniel, 

The WWPN listed in your Q FCP outputs and that from the SCSIDISC tool are 
similar enough that I suspect they are different subchannels on the same 
FCP card, and not of your target DASD subsystem.  Are there other WWPNs 
listed in the SCSIDISC output beyond what you have sent?  If so, what does 
SCSIDISC see with them?  If not, has the switch been told about the NPIV 
WWPNs that the VM LPAR will be using (that 

Re: Channel Contention

[Raymond Higgs]

The IBM z/VM Operating System  wrote on 
02/18/2010 10:48:52 AM:

> Brian Nielsen  
> Sent by: The IBM z/VM Operating System 
> 
> 02/18/2010 10:48 AM
> 
> Please respond to
> The IBM z/VM Operating System 
> 
> To
> 
> IBMVM@LISTSERV.UARK.EDU
> 
> cc
> 
> Subject
> 
> Re: Channel Contention
> 
> Being that this is a one-time effort, I wouldn't worry too much about 
> where the bottleneck is because you indicate can't do anything about it 
> 
> anyway.  Better is simply to estimate the amount of time it will take to 

> 
> some gross level of precision.  If you were going to be doing it 
regularl
> y 
> then it becomes a more important question and could drive configuration 
> 
> changes.
> 
> I know tape drives are not involved in your process, but I do know that 
> 
> DDR is very effective at driving the channel to them.  We have a bank of 

> 4 
> 3590's shared between VM and z/OS on 2 ESCON channels.  When I had 2 
DDR'
> s 
> writing to 2 of the tapes drives it effectivley saturated the channels 
> 
> from the VM LPAR and the z/OS jobs trying to use the other 2 tape drives 

> 
> suffered horribly.  The reports from Velocity's ESAMAP made it easy to 
> 
> diagnose this cross-LPAR interference.  (I ended up moving my DDR's to a 

> 
> different time slot.)  I mention this to say that it will be good that 
> 
> there will not be other production workload going on to the DASD when 
you
> 
> do this.
> 
> BTW, the 3-4% channel utilization I mentioned in my other post is on 2G 
> 
> FICON channels.  Obviously, 4G or 8G channels would make a big 
difference
> .
> 
> Brian Nielsen

I wouldn't rely on those System Activity Display numbers with big IOs to 
determine channel capacity.  That 3-4% is how busy the powerpc processor 
is in the channel.  With big IOs, it's just idling, waiting on the DMA 
engines.  If you add more IOs to the channel, you'll max out some piece of 
hardware way before the SAD panel shows 50%.  With one of my 8 gig FCP 
channels, I see 25% at max read bandwidth.  I think ficon would yield 
similar numbers.

The processor in 2 and 4 gig channels is the same.  The processor in 8 gig 
channels is quite a bit faster.

Ray Higgs
System z FCP Development
Bld. 706, B24
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
rayhi...@us.ibm.com

Re: FCP Redbooks

[Raymond Higgs]

The IBM z/VM Operating System  wrote on 
12/17/2009 09:17:11 PM:

> "O'Brien, Dennis L"  
> Sent by: The IBM z/VM Operating System 
> 
> 12/17/2009 09:17 PM
> 
> Please respond to
> The IBM z/VM Operating System 
> 
> To
> 
> IBMVM@LISTSERV.UARK.EDU
> 
> cc
> 
> Subject
> 
> FCP Redbooks
> 
> Our storage architect asked me about using FCP channels for Linux on
> z/VM.  Are there any good Redbooks or other documentation that's 
> recent?  He couldn't find anything that mentioned z10.  He's 
> particularly interested in limits (guests/LUN's/device addresses per
> channel, etc).  CP Planning and Administration discusses 
> configuration, but doesn't mention limits.  I did find a discussion 
> of limits in the Introducing N_Port Identifier Virtualization for 
> IBM System z9 Redbook paper.  That paper is from 2006.  Has anything
> changed since then?  I also found Steve Wilkins' presentation Using 
> z/VM in a SCSI Environment.  It has some good recommendations.
>   
>   Dennis
> 
> What's the difference between an Escalade and a golf ball?  Tiger 
> Woods can drive a golf ball 400 yards.


Dennis,

I can only speak for the channel.  z/VM, and Linux may have their own 
limits that conflict with ours.

Fred's Intro. Redpaper is still accurate.  The number of guests, and 
number of device addresses per channel is really the same question.  The 
absolute max is 480.  If you are using NPIV, then our recommendation is 
32.  The number of target ports per channel that we recommend is no more 
than 128, though I think the real max is 510.  The number of luns per 
channel is 4096 max.

If you navigate from this hints webpage:

http://www.ibm.com/developerworks/linux/linux390/distribution_hints.html

You can find "How to use FC-attached SCSI devices with Linux on System z", 
and "Device Drivers, Features, and Commands" for the distro that you're 
interested.  Both have a lot of information on how to use FCP.

I hope this helps!

Ray Higgs
System z FCP Development
Bld. 706, B24
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
rayhi...@us.ibm.com

Re: OT: eclipse based info center

[Raymond Higgs]

The IBM z/VM Operating System  wrote on 
09/16/2009 04:31:55 PM:

> Rob van der Heij  
> Sent by: The IBM z/VM Operating System 
> 
> 09/16/2009 04:31 PM
> 
> Please respond to
> The IBM z/VM Operating System 
> 
> To
> 
> IBMVM@LISTSERV.UARK.EDU
> 
> cc
> 
> Subject
> 
> OT: eclipse based info center
> 
> Is it time already for another rant on "the new way to reading books" 
;-)
> 
> I had to look up stuff in one of the z/OS books and landed on the
> "z/OS V1R9 Information Center" - horrible experience, apart from the
> content. When you select something on the menu the contents is
> refreshed again and moves under the cursor, causing it to hover over a
> completely different section and pop up with stuff. And dog slow.
> Takes me 10 seconds to find the selected section has just a single
> sentence and no content at all. And I thought the option to "print
> topic and subtopics" would be useful... guess what, that's also full
> of links and navigation stuff. What are these people thinking?
> 
> But I've found the PDF - it downloads in less then a minute and after
> that I can browse it without taking a class on how to read IBM
> documentation.
> 
> Rob

Re: FCP using z/VM 5.3 and RedHat Linux 4.6

[Raymond Higgs]

The IBM z/VM Operating System  wrote on 
05/26/2008 03:46:17 PM:

> Martin, Terry R. (CMS/CTR) (CTR) wrote:
> >
> > 
> >
> > 
> >
> > e49l124d kernel: zfcp: error: Point-to-point fibrechannel 
> > configuration detected at adapter 0.0.5500 unsupported, shutting down 
> > adapter
> >
> Sounds like you may not have support built into the particular Linux 
> kernel you're running. Tried a kernel reconfig?
> 
> -- 
> Jack J. Woehr# "Self-delusion is
> http://www.well.com/~jax #  half the battle!"
> http://www.softwoehr.com #  - Zippy the Pinhead

Support for switched topologies came first.  If you want point-to-point, 
you have to use a newer version of linux.  According to one of the cross 
reference sites, you need at least 2.6.12.

Ray Higgs
System z FCP Development
Bld. 706, B24
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
[EMAIL PROTECTED]

Re: reasons to NOT use EDEV

[Raymond Higgs]

The IBM z/VM Operating System  wrote on 
02/19/2008 04:53:22 PM:

> Given that performance is very nearly the same between SAN and EDEV 
> connections, or acceptably so, then I would recommend that you go with 
> the EDEV approach. Just have your SAN folks carve out a big chunk of SAN 

> storage for your use, and then allocate EDEV (FBA type) disk storage for 

> guests as needed from that large pool. You want to minimize your 
> contacts with the SAN folks, I think.
> 
> Good luck.
> 

I don't think the performance is very close.  Here's some anecdotal 
evidence...

For Danu (z9) GA 3, we (firmware) were roped into measuring the FBA 
performance difference: GA 2 vs GA 3.  We wanted to determine if firmware 
changes that helped direct attached IOs/sec also helped FBA.  The testing 
procedure was ad hoc.  We found a configuration of FBA devices that would 
get the max 4k SCSI reads/sec from 1 channel.  We got numbers with the GA 
2 firmware.  We upgraded the FCP firmware to GA 3, and used the same FBA 
config to get a new set of numbers.  We then compared to direct attached 
4k SCSI reads/sec numbers that we already had.  Direct attached was 
something like 50% faster for both levels of firmware.

Sorry that I don't have more details, but it was over a year ago, and I no 
longer have the string of notes.

If anyone out there is contemplating FBA vs direct attached and 
performance is a concern, I think you should do some of your own tests 
before deciding.

Ray Higgs
System z FCP Development
Bld. 706, B24
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
[EMAIL PROTECTED]

Re: CMS TSM admin client

[Raymond Higgs]

The IBM z/VM Operating System  wrote on 
12/28/2007 09:40:22 AM:

> On Thursday, 12/27/2007 at 04:18 EST, "McKown, John" 
> <[EMAIL PROTECTED]> wrote:
> 
> > How about if IBM allowed a zIIP or zAAP to do this?
> 
> zIIPs and zAAPs are designed to reduce the software costs associated 
with 
> z/OS.  You could get an IFL for the same price and use any left-over 
> cycles for Other Work.
> 
> But I think I'd rather see the System z FCP adapter support being a 
> *target* WWPN+LUN.  That would open up the moral equivalent of SAN 
Volume 
> Controller (or maybe SVC itself!) on System z. The targeted device 
driver 
> could emulate any kind of SCSI device it wanted to, hiding the backend 
> implementation, just as SVC does.  You'd get the benefits of emulation 
> without the overhead of iSCSI.
> 
> Alan Altmark
> z/VM Development
> IBM Endicott


We already have firmware to make a channel a target port.  Our STV (SCSI 
test vehicle) emulates SCSI disks.  There's no reason why it couldn't be 
extended to function like an SVC.

Ray Higgs
System z FCP Development
Bld. 706, B24
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
[EMAIL PROTECTED]

Re: SAN and Linux on zSeries

[Raymond Higgs]

The IBM z/VM Operating System  wrote on 
12/10/2007 03:08:00 PM:

> Now, if only all the storage controllers out there would implement SMI-S 

> interfaces...
> 
> Alan Altmark
> z/VM Development
> IBM Endicott

We can't hide behind that excuse.  A quick google shows IBM storage, EMC, 
Hitachi, Storage Tek, Cisco and Brocade have some level of support :)

Ray Higgs
System z FCP Development
Bld. 706, B24
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
[EMAIL PROTECTED]

Re: SAN and Linux on zSeries

[Raymond Higgs]

The IBM z/VM Operating System  wrote on 
12/10/2007 10:50:40 AM:

> > SMI-S provides a standard management API to make tool building easier. 

> 
> This is an interesting statement. Would there be interest in a CMS-
> based SMI-S library to enable building tools to automate this sort 
> of stuff, and would IBM integrate it if one was available? Seems to 
> be a good enablement project if it were.
> 
> 

There would probably need to be some i390 and SE work done too.  They 
manage the WWPNs, and really have the complete picture.

Ray Higgs
System z FCP Development
Bld. 706, B24
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
[EMAIL PROTECTED]

Re: SAN and Linux on zSeries

[Raymond Higgs]

The IBM z/VM Operating System  wrote on 
12/10/2007 08:33:38 AM:

> Ray,
> You mention number of usable NPIV subchannels are subject to the 
> switch limitations.
> 
> Are switch limitations what the IBM Redbook is referencing when it 
> says don't "use more than 32 subchannels per physical channel in NPIV 
mode"? 
> With the Brocade switch I could use 255 subchannels in NPIV mode?
> 
> (see "Introducing N_Port Identifier Virtualization
> for IBM System z9"  page 4
> http://www.redbooks.ibm.com/redpapers/pdfs/redp4125.pdf )
> 
> " Configuration considerations 
> Some general recommendations for using NPIV include:
> Do not use more than 32 subchannels per physical channel in NPIV mode.
>  Also, do not perform more than 128 total target logins (for 
> example, in a configuration with 32 subchannels, limit the number of
> target logins to no more than an average of 4). Using more 
> subchannels, target logins, or both can create timeouts."
> 
> 

John,

Yes, this is partially a switch limitation.  The process to log an NPIV 
port into the fabric takes milliseconds.  When there are lots of these, 
the time becomes seconds.  When we did NPIV bringup, we discovered 
situations where linux would time out and escalate its recovery.  The 
easiest way to cause this is with fibre pulls.  Though, any event that 
causes link down, up should lead to it too.

FWIW, using lots of small zones may help alleviate these operating system 
time outs.  When the link goes down/up, the number of state change 
notifications generated by the the switch will be less.  So the switch 
will have more resources to handle logins.

If you are brave, you can try to use more NPIV subchannels.  There is 
nothing in the FCP firmware to prevent you from doing so.

Keep in mind that when we make configuration recommendations, we have to 
make them for the worst case.

It's nice to hear that people are actually interested :)

Ray Higgs
System z FCP Development
Bld. 706, B24
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
[EMAIL PROTECTED]

Re: SAN and Linux on zSeries

[Raymond Higgs]

The IBM z/VM Operating System  wrote on 
12/08/2007 07:31:05 PM:

> > > Handing each guest its own HBA (host bus adapter,
> > > the open systems term for and FCP adapter) kind of blows
> > > one of the reasons to go virtual.
> >
> > Eh?  480 servers can use a single FCP adapter (chpid) concurrently. 
That's
> > the whole point of N_port ID virtualization: 480 separate fabric
> > endpoints.
> 
> No no ... it's not a question of how many guests can share an
> FCP adapter.  The trouble is that directly connected guests are
> managerially more like discrete systems, so there is no way from VM
> to manage the storage.
> 
> N_port ID virtualization adds to this because the storage
> is restricted to a specific FCP adapter.  Then if we need to
> give that LUN (or LUNs) to a different guest, we have to give
> the FCP adapter to it.  What if the recipient already has an FCP
> adapter?  It might be better to leverage the adapter already there.
> 
> -- R;

Well, NPIV was necessary.  Without it, lots of functions were broken.  The 
LUN access control stuff that we implemented only fixed some of these 
problems.

FWIW, the problems you describe aren't unique to the mainframe.  SMI-S is 
supposed to be the industry standard fix for all of the management 
problems, but I do not know much about it.

Ray Higgs
System z FCP Development
Bld. 706, B24
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
[EMAIL PROTECTED]

Re: SAN and Linux on zSeries

[Raymond Higgs]

The IBM z/VM Operating System  wrote on 
12/07/2007 11:25:06 AM:

> On Friday, 12/07/2007 at 06:16 EST, Rick Troth <[EMAIL PROTECTED]> wrote:
> 
> > Handing each guest its own HBA (host bus adapter,
> > the open systems term for and FCP adapter) kind of blows
> > one of the reasons to go virtual.
> 
> Eh?  480 servers can use a single FCP adapter (chpid) concurrently. 
That's 
> the whole point of N_port ID virtualization: 480 separate fabric 
> endpoints.
> 
> Alan Altmark
> z/VM Development
> IBM Endicott

Not all 480 subchannels can be used with NPIV because of switch 
limitations.  The 480 number was chosen arbitrarily from the layout of the 
SIGA vector long before NPIV was implemented.  I think Brocade supports 
the most virtual nports per physical port, 255.

Ray Higgs
System z FCP Development
Bld. 706, B24
2455 South Road
Poughkeepsie, NY 12601
(845) 435-8666,  T/L 295-8666
[EMAIL PROTECTED]