Re: information on the config option -- node.session.iscsi.FastAbort = No
On 04/28/2010 12:43 PM, Mike Christie wrote: On 04/28/2010 10:40 AM, maguar887 wrote: We are currently running open iscsi version 2.0-871 on RHEL 5.3 (2.6.18-92.1.6.0.2.el5) against a Dell Equallogic iScsi SAN group (firmware 4.3.5) You need to upgrade your kernel. It had a bug with eql targets in the async logout path. It is fixed in 5.4 and 5.5 kernels. Actually it was fixed in 5.3 too. The weird thing here is that we get this error *before* we get the async logout request. Do you have multiple sessions? Is sdt accessed through session 22 (run iscsiadm -m session -P 3 to see) or a different session? If a different session is there more log and can you send it? Apr 18 11:18:11 serv02 kernel: sd 24:0:0:0: SCSI error: return code = 0x0007 Apr 18 11:18:11 serv02 kernel: end_request: I/O error, dev sdt, sector 27424 Apr 18 11:18:11 serv02 kernel: sd 24:0:0:0: SCSI error: return code = 0x0007 Apr 18 11:18:11 serv02 kernel: end_request: I/O error, dev sdt, sector 27520 Apr 18 11:18:11 serv02 kernel: Buffer I/O error on device sdt1, logical block 3436 Apr 18 11:18:11 serv02 kernel: lost page write due to I/O error on sdt1 Apr 18 11:18:11 serv02 kernel: Aborting journal on device sdt1. Apr 18 11:18:11 serv02 kernel: ext3_abort called. Apr 18 11:18:11 serv02 kernel: EXT3-fs error (device sdt1): ext3_journal_start_sb: Detected aborted journal Apr 18 11:18:11 serv02 kernel: Remounting filesystem read-only Apr 18 11:18:12 serv02 iscsid: Target requests logout within 3 seconds for connection Apr 18 11:18:16 serv02 iscsid: connection22:0 is operational after recovery (1 attempts) -- 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: information on the config option -- node.session.iscsi.FastAbort = No
On 04/28/2010 12:56 PM, Mike Christie wrote: On 04/28/2010 12:43 PM, Mike Christie wrote: On 04/28/2010 10:40 AM, maguar887 wrote: We are currently running open iscsi version 2.0-871 on RHEL 5.3 (2.6.18-92.1.6.0.2.el5) against a Dell Equallogic iScsi SAN group (firmware 4.3.5) You need to upgrade your kernel. It had a bug with eql targets in the async logout path. It is fixed in 5.4 and 5.5 kernels. Actually it was fixed in 5.3 too. The weird thing here is that we get this error *before* we get the async logout request. Do you have multiple sessions? Is sdt accessed through session 22 (run iscsiadm -m session -P 3 to see) or a different session? If a different session is there more log and can you send it? Could you also take a ethereal trace? Looking at the iscsi code for that kernel the only place to get 0x0007 seems to be this: if (rhdr-response != ISCSI_STATUS_CMD_COMPLETED) { sc-result = DID_ERROR 16; goto out; } which means the target did not complete the command. The initiator/scsi-layer would have retried this command up to 5 times before failing. -- 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: information on the config option -- node.session.iscsi.FastAbort = No
On 04/28/2010 02:07 PM, maguar887 wrote: Mike, Thanks for the info! Do you know exactly which kernel it was patched in? what is available to us is: 2.6.18-194.0.0.0.4.el5Matt Forget the upgrade your kernel comment. The bug I was thinking about was fixed in the kernel you were using, 2.6.18-92.1.6.0.2.el5 (that is actually the kernel for RHEL 5.2 but you said you were using 5.3 btw). Could you get me a ethereal trace, so I can see if the target is failing the IO on us? Matt On Apr 28, 1:58 pm, Mike Christiemicha...@cs.wisc.edu wrote: On 04/28/2010 12:56 PM, Mike Christie wrote: On 04/28/2010 12:43 PM, Mike Christie wrote: On 04/28/2010 10:40 AM, maguar887 wrote: We are currently running open iscsi version 2.0-871 on RHEL 5.3 (2.6.18-92.1.6.0.2.el5) against a Dell Equallogic iScsi SAN group (firmware 4.3.5) You need to upgrade your kernel. It had a bug with eql targets in the async logout path. It is fixed in 5.4 and 5.5 kernels. Actually it was fixed in 5.3 too. The weird thing here is that we get this error *before* we get the async logout request. Do you have multiple sessions? Is sdt accessed through session 22 (run iscsiadm -m session -P 3 to see) or a different session? If a different session is there more log and can you send it? Could you also take a ethereal trace? Looking at the iscsi code for that kernel the only place to get 0x0007 seems to be this: if (rhdr-response != ISCSI_STATUS_CMD_COMPLETED) { sc-result = DID_ERROR 16; goto out; } which means the target did not complete the command. The initiator/scsi-layer would have retried this command up to 5 times before failing. -- 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 athttp://groups.google.com/group/open-iscsi?hl=en.- Hide quoted text - - Show quoted text - -- 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: information on the config option -- node.session.iscsi.FastAbort = No
Sorry for the confusion, we had 2 seperate systems, and I pulled the onfo from the wrong one. The system with the issue turns out to be RHEL 5.2 with this kernel: 2.6.18-53.1.19.0.1.el5 I'll try and reproduce and get a packet capture On Apr 28, 3:28 pm, Mike Christie micha...@cs.wisc.edu wrote: On 04/28/2010 02:07 PM, maguar887 wrote: Mike, Thanks for the info! Do you know exactly which kernel it was patched in? what is available to us is: 2.6.18-194.0.0.0.4.el5Matt Forget the upgrade your kernel comment. The bug I was thinking about was fixed in the kernel you were using, 2.6.18-92.1.6.0.2.el5 (that is actually the kernel for RHEL 5.2 but you said you were using 5.3 btw). Could you get me a ethereal trace, so I can see if the target is failing the IO on us? Matt On Apr 28, 1:58 pm, Mike Christiemicha...@cs.wisc.edu wrote: On 04/28/2010 12:56 PM, Mike Christie wrote: On 04/28/2010 12:43 PM, Mike Christie wrote: On 04/28/2010 10:40 AM, maguar887 wrote: We are currently running open iscsi version 2.0-871 on RHEL 5.3 (2.6.18-92.1.6.0.2.el5) against a Dell Equallogic iScsi SAN group (firmware 4.3.5) You need to upgrade your kernel. It had a bug with eql targets in the async logout path. It is fixed in 5.4 and 5.5 kernels. Actually it was fixed in 5.3 too. The weird thing here is that we get this error *before* we get the async logout request. Do you have multiple sessions? Is sdt accessed through session 22 (run iscsiadm -m session -P 3 to see) or a different session? If a different session is there more log and can you send it? Could you also take a ethereal trace? Looking at the iscsi code for that kernel the only place to get 0x0007 seems to be this: if (rhdr-response != ISCSI_STATUS_CMD_COMPLETED) { sc-result = DID_ERROR 16; goto out; } which means the target did not complete the command. The initiator/scsi-layer would have retried this command up to 5 times before failing. -- 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 athttp://groups.google.com/group/open-iscsi?hl=en.-Hide quoted text - - Show quoted text - -- 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 athttp://groups.google.com/group/open-iscsi?hl=en.- Hide quoted text - - Show quoted text - -- 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.