Re: open-iscsi fails on SLES 10 SP3 kernel but works on SP2 kernel.
On 18 May 2010 at 11:14, Anton Altaparmakov wrote: Hi, We upgraded from SLES 10 SP2 (x86_64) to SLES 10 SP3 (x86_64) which caused our open-iscsi to iscsitarget connections to fail. Downgrading to the SP2 kernel makes it work again. Using the SP3 kernel running we see: Did you just install the SP§, or did you install SP3 + Updates? (Just being curious) [...] Regards, Ulrich -- You received this message because you are subscribed to the Google Groups open-iscsi group. To post to this group, send email to open-is...@googlegroups.com. To unsubscribe from this group, send email to open-iscsi+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/open-iscsi?hl=en.
Re: open-iscsi fails on SLES 10 SP3 kernel but works on SP2 kernel.
Hi Ulrich, On 19 May 2010, at 07:35, Ulrich Windl wrote: On 18 May 2010 at 11:14, Anton Altaparmakov wrote: We upgraded from SLES 10 SP2 (x86_64) to SLES 10 SP3 (x86_64) which caused our open-iscsi to iscsitarget connections to fail. Downgrading to the SP2 kernel makes it work again. Using the SP3 kernel running we see: Did you just install the SP§, or did you install SP3 + Updates? (Just being curious) + Updates. And I have been trying every kernel update (for about two months now) and it always fails so I have had to downgrade back to the SP2 kernel each time... I guess I ought to try a vanilla SP3 kernel at some point to see if that makes it work. It is a production system so doing testing is very difficult when reboots are needed... Best regards, Anton -- Anton Altaparmakov aia21 at cam.ac.uk (replace at with @) Unix Support, Computing Service, University of Cambridge, CB2 3QH, UK Linux NTFS maintainer, http://www.linux-ntfs.org/ -- You received this message because you are subscribed to the Google Groups open-iscsi group. To post to this group, send email to open-is...@googlegroups.com. To unsubscribe from this group, send email to open-iscsi+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/open-iscsi?hl=en.
SCSI reservation and Suse 10.2
We are using suse 10.2 open-iscsi version 2.0-707 (shipped version) We are sharing the iscsi device with vmware 4.0 esx server. What we see happen is that we get io error from a scsi reservation, we wait 100ms and re-try and we always get the reservation error, we return the error. When the vmware vixdix library sees an eio closes and re-opens then seeks to block zero and reads which fails and then the operation fails We captured the the reservation error and waited 10 seconds and then retured the error to to the library which closed, open, seeked, read successfully and continued. Questions: Is there a new version of the initiator which we should upgrade to to improve this situation. Are there setting for the initiator that can help -- You received this message because you are subscribed to the Google Groups open-iscsi group. To post to this group, send email to open-is...@googlegroups.com. To unsubscribe from this group, send email to open-iscsi+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/open-iscsi?hl=en.
Re: open-iscsi fails on SLES 10 SP3 kernel but works on SP2 kernel.
On 05/18/2010 05:14 AM, Anton Altaparmakov wrote: # iscsiadm -m session -P3 iSCSI Transport Class version 2.0-724 iscsiadm version 2.0-868 Target: iqn.2008-02.xenstorenet.nessie:arup301 Current Portal: 10.10.10.1:3260,1 Persistent Portal: 10.10.10.1:3260,1 ** Interface: ** Iface Name: default Iface Transport: tcp Iface Initiatorname: iqn.2008-02.xenstorenet.hydra:redstone Iface IPaddress: 10.10.10.2 Iface HWaddress: default Iface Netdev: default SID: 5 iSCSI Connection State: LOGGED IN iSCSI Session State: Unknown Internal iscsid Session State: NO CHANGE Negotiated iSCSI params: HeaderDigest: CRC32C DataDigest: CRC32C You might need this fix: http://git.kernel.org/?p=linux/kernel/git/jejb/scsi-misc-2.6.git;a=commit;h=96b1f96dcab87756c0a1e7ba76bc5dc2add82b88 -- You received this message because you are subscribed to the Google Groups open-iscsi group. To post to this group, send email to open-is...@googlegroups.com. To unsubscribe from this group, send email to open-iscsi+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/open-iscsi?hl=en.
Re: iscsid: Kernel reported iSCSI connection 1:0 error (1020) state (3)
On 05/17/2010 07:39 PM, 立凡 王 wrote: The version of the kernel is '2.6.32.12-114.fc12.i686.PAE'. The log file after debugging turned on is - http://people.chu.edu.tw/~b8902110/temp/messages0517 There are lots of below messages in the log. kernel: session1: iscsi_eh_cmd_timed_out scsi cmd f5eb8480 timedout kernel: session1: iscsi_eh_cmd_timed_out return timer reset And I didn't see messages like aborting sc 0x1234432 Ah, I guess that is good news and bad news. It is good news we are avoiding starting the scsi eh when we do not have to. However it is bad news because the iscsi_eh_cmd_timed_out messages means it is taking the target a long time to process the IO. What is the value of your scsi cmd timeout? cat /sys/block/sdX/device/timeout And this is connected to that HP target, right? -- You received this message because you are subscribed to the Google Groups open-iscsi group. To post to this group, send email to open-is...@googlegroups.com. To unsubscribe from this group, send email to open-iscsi+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/open-iscsi?hl=en.
Re: SCSI reservation and Suse 10.2
On 05/18/2010 05:36 PM, JimP wrote: We are using suse 10.2 open-iscsi version 2.0-707 (shipped version) We are sharing the iscsi device with vmware 4.0 esx server. What we see happen is that we get io error from a scsi reservation, we wait 100ms and re-try and we always get the reservation error, we return the error. When the vmware vixdix library sees an eio closes and re-opens then seeks to block zero and reads which fails and then the operation fails We captured the the reservation error and waited 10 seconds and then retured the error to to the library which closed, open, seeked, read successfully and continued. Questions: Is there a new version of the initiator which we should upgrade to to improve this situation. Are there setting for the initiator that can help The iscsi initiator does not do anything when it comes to scsi reservations. The scsi layer and apps above it handle it. The iscsi layer is just a passthrough for those requests. Do you see any scsi error messages in /var/log/messages? -- You received this message because you are subscribed to the Google Groups open-iscsi group. To post to this group, send email to open-is...@googlegroups.com. To unsubscribe from this group, send email to open-iscsi+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/open-iscsi?hl=en.
Q: syslog frequency for login rejected
Hi, I have a question: During a hardware fault at the disk system (which is external to the iSCSI target) a SLES10 SP2+Updates (x86_64) server reported login errors like follows every few seconds. I wonder if the current version of the software still does that. As the hardware fault took significant time to repair, the syslog was flooded significantly with those messages. Ulrich --- [...] Apr 19 17:10:19 host3 iscsid: conn 0 login rejected: target error (03/01) Apr 19 17:10:20 host3 iscsid: conn 0 login rejected: target error (03/01) Apr 19 17:10:20 host3 iscsid: conn 0 login rejected: target error (03/01) Apr 19 17:10:20 host3 iscsid: conn 0 login rejected: target error (03/01) Apr 19 17:10:20 host3 iscsid: conn 0 login rejected: target error (03/01) Apr 19 17:10:20 host3 iscsid: conn 0 login rejected: target error (03/01) Apr 19 17:10:20 host3 iscsid: conn 0 login rejected: target error (03/01) Apr 19 17:10:20 host3 iscsid: conn 0 login rejected: target error (03/01) Apr 19 17:10:20 host3 iscsid: conn 0 login rejected: target error (03/01) Apr 19 17:10:21 host3 iscsid: conn 0 login rejected: target error (03/01) Apr 19 17:10:21 host3 iscsid: conn 0 login rejected: target error (03/01) Apr 19 17:10:21 host3 iscsid: conn 0 login rejected: target error (03/01) Apr 19 17:10:21 host3 iscsid: conn 0 login rejected: target error (03/01) Apr 19 17:10:21 host3 iscsid: conn 0 login rejected: target error (03/01) Apr 19 17:10:21 host3 iscsid: conn 0 login rejected: target error (03/01) Apr 19 17:10:21 host3 iscsid: conn 0 login rejected: target error (03/01) Apr 19 17:10:21 host3 iscsid: conn 0 login rejected: target error (03/01) Apr 19 17:10:22 host3 kernel: session2: iscsi: session recovery timed out after 120 secs Apr 19 17:10:22 host3 kernel: session24: iscsi: session recovery timed out after 120 secs [...] -- You received this message because you are subscribed to the Google Groups open-iscsi group. To post to this group, send email to open-is...@googlegroups.com. To unsubscribe from this group, send email to open-iscsi+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/open-iscsi?hl=en.
Re: Q: syslog frequency for login rejected
On 05/19/2010 08:43 AM, Ulrich Windl wrote: Hi, I have a question: During a hardware fault at the disk system (which is external to the iSCSI target) a SLES10 SP2+Updates (x86_64) server reported login errors like follows every few seconds. I wonder if the current version of the software still does that. Was that the first bug you posted about on this list? Yeah, it still does that. I think I said on the list before that I was going to fix it in the next release but did not get time and then Erez was working on other logging improvements so I was hoping to finish his work first. -- You received this message because you are subscribed to the Google Groups open-iscsi group. To post to this group, send email to open-is...@googlegroups.com. To unsubscribe from this group, send email to open-iscsi+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/open-iscsi?hl=en.
Re: open-iscsi fails on SLES 10 SP3 kernel but works on SP2 kernel.
On 05/19/2010 09:20 AM, Anton Altaparmakov wrote: Hi Mike, On 19 May 2010, at 12:48, Mike Christie wrote: On 05/18/2010 05:14 AM, Anton Altaparmakov wrote: # iscsiadm -m session -P3 iSCSI Transport Class version 2.0-724 iscsiadm version 2.0-868 Target: iqn.2008-02.xenstorenet.nessie:arup301 Current Portal: 10.10.10.1:3260,1 Persistent Portal: 10.10.10.1:3260,1 ** Interface: ** Iface Name: default Iface Transport: tcp Iface Initiatorname: iqn.2008-02.xenstorenet.hydra:redstone Iface IPaddress: 10.10.10.2 Iface HWaddress: default Iface Netdev: default SID: 5 iSCSI Connection State: LOGGED IN iSCSI Session State: Unknown Internal iscsid Session State: NO CHANGE Negotiated iSCSI params: HeaderDigest: CRC32C DataDigest: CRC32C You might need this fix: http://git.kernel.org/?p=linux/kernel/git/jejb/scsi-misc-2.6.git;a=commit;h=96b1f96dcab87756c0a1e7ba76bc5dc2add82b88 Thank you for the suggestion! I just tried it by changing that line, recompiling and replacing the SLES 10 libiscsi.ko module with the newly compiled one, then stopped the open-iscsi service and ensured the module was unloaded and then tried again, this time it loaded the new module (and complained it wasn't supported by Novell so definitely loading the new one) and it still timed out the same way. )-: Any other ideas? Is there anyway you can give me the libiscsi.c, iscsi_tcp.c and if it exists the libiscsi_tcp.c files for kernel that worked and did not work? Or could you also take a ethereal trace for when it worked and when it did not? -- You received this message because you are subscribed to the Google Groups open-iscsi group. To post to this group, send email to open-is...@googlegroups.com. To unsubscribe from this group, send email to open-iscsi+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/open-iscsi?hl=en.
questions about iscsi_transport hooks
Mike, I am working on the attached patch sent by you for allowing user space to request session. This patch is created on top of the older iscsi_transport. In the latest iscsi_transport, create_session hook requires struct iscsi_endpoint to be passed as function parameter. To port this patch on latest iscsi_transport, looks like I need to add additional driver hooks: ep_connect, ep_poll, and ep_disconnect. I want to understand what low level drivers like qla4xxx which operates in session mode are supposed to do in ep_* hooks, and how ep_* hooks are different than create_conn, start_conn, and stop_conn Looking forward for your suggestions and reply. Thanks, Vikas. -- You received this message because you are subscribed to the Google Groups open-iscsi group. To post to this group, send email to open-is...@googlegroups.com. To unsubscribe from this group, send email to open-iscsi+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/open-iscsi?hl=en. 0001-qla4xxx-allow-user-space-to-request-session-creation.patch Description: 0001-qla4xxx-allow-user-space-to-request-session-creation.patch
Re: open-iscsi fails on SLES 10 SP3 kernel but works on SP2 kernel.
Hi Mike, [Note: Attachment sent off list so as not to pollute everyone's inbox.] On 19 May 2010, at 20:14, Mike Christie wrote: On 05/19/2010 09:20 AM, Anton Altaparmakov wrote: Hi Mike, On 19 May 2010, at 12:48, Mike Christie wrote: On 05/18/2010 05:14 AM, Anton Altaparmakov wrote: # iscsiadm -m session -P3 iSCSI Transport Class version 2.0-724 iscsiadm version 2.0-868 Target: iqn.2008-02.xenstorenet.nessie:arup301 Current Portal: 10.10.10.1:3260,1 Persistent Portal: 10.10.10.1:3260,1 ** Interface: ** Iface Name: default Iface Transport: tcp Iface Initiatorname: iqn.2008-02.xenstorenet.hydra:redstone Iface IPaddress: 10.10.10.2 Iface HWaddress: default Iface Netdev: default SID: 5 iSCSI Connection State: LOGGED IN iSCSI Session State: Unknown Internal iscsid Session State: NO CHANGE Negotiated iSCSI params: HeaderDigest: CRC32C DataDigest: CRC32C You might need this fix: http://git.kernel.org/?p=linux/kernel/git/jejb/scsi-misc-2.6.git;a=commit;h=96b1f96dcab87756c0a1e7ba76bc5dc2add82b88 Thank you for the suggestion! I just tried it by changing that line, recompiling and replacing the SLES 10 libiscsi.ko module with the newly compiled one, then stopped the open-iscsi service and ensured the module was unloaded and then tried again, this time it loaded the new module (and complained it wasn't supported by Novell so definitely loading the new one) and it still timed out the same way. )-: Any other ideas? Is there anyway you can give me the libiscsi.c, iscsi_tcp.c and if it exists the libiscsi_tcp.c files for kernel that worked and did not work? Sure, attached... The sp2-iscsi directory is the working one and the sp3-iscsi directory is the non-working one (note it includes the patch that you referenced above, the original SP3 code did not have that). Or could you also take a ethereal trace for when it worked and when it did not? Not sure about ethereal. Would tcpdump do? What options would you like me to run it with? Thanks a lot for your help! Best regards, Anton -- Anton Altaparmakov aia21 at cam.ac.uk (replace at with @) Unix Support, Computing Service, University of Cambridge, CB2 3QH, UK Linux NTFS maintainer, http://www.linux-ntfs.org/ -- You received this message because you are subscribed to the Google Groups open-iscsi group. To post to this group, send email to open-is...@googlegroups.com. To unsubscribe from this group, send email to open-iscsi+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/open-iscsi?hl=en.