On 03/11/2010 07:25 AM, Oriol Morell wrote:
Mike,

in the var-log-messages we can see the following error but not all the time:

     589125 Mar 11 12:39:48 kabuki07 kernel: connection1:0: detected conn error
     (1020)

This (1020) just means the target dropped the connection on us.



     589126 Mar 11 12:39:49 kabuki07 iscsid: Kernel reported iSCSI connection 
1:0
     error (1020) state (3)
     589127 Mar 11 12:39:52 kabuki07 iscsid: connection1:0 is operational after
     recovery (1 attempts)


We were able to reconnect right way.


     589128 Mar 11 12:40:35 kabuki07 kernel: INFO: task async/0:2554 blocked for
     more than 120 seconds.

This means it might have been happening a lot because some IO has not completed for over two minutes......



     589129 Mar 11 12:40:35 kabuki07 kernel: "echo 0>
     /proc/sys/kernel/hung_task_timeout_secs" disables this message.


.......

     kernel_thread_helper+0x0/0x10*
     589161 Mar 11 12:40:43 kabuki07 kernel: connection1:0: detected conn error
     (1020)

Target dropped connection on us a again.


     589162 Mar 11 12:40:44 kabuki07 kernel: sd 1:0:0:0: timing out command,
     waited 180s

Yeah, it looks like we have been retrying this command a couple times and the problem kept happening. The scsi layer eventually says enough and fails the command after retrying for 180 secs. So this is also why you see that stack trace and the message about the blocked task for more than 120 secs.



     589163 Mar 11 12:40:44 kabuki07 kernel: sd 1:0:0:0: [sdb] Unhandled error 
code
     589164 Mar 11 12:40:44 kabuki07 kernel: sd 1:0:0:0: [sdb] Result:
     hostbyte=DID_TRANSPORT_DISRUPTED driverbyte=DRIVER_OK
     589165 Mar 11 12:40:44 kabuki07 kernel: sd 1:0:0:0: [sdb] CDB: Read(10): 28
     00 00 00 00 00 00 00 08 00
     589166 Mar 11 12:40:44 kabuki07 kernel: end_request: I/O error, dev sdb,
     sector 0

Cmd is failed.

     589167 Mar 11 12:40:44 kabuki07 kernel: Buffer I/O error on device sdb,
     logical block 0
     589168 Mar 11 12:40:44 kabuki07 kernel: unable to read partition table


It seems like the target is not liking something we are doing.

The target logs do not help me much. I looked through them, but I am not sure what they are telling me. The storagetek guys would need to decipher them for us.

--
You received this message because you are subscribed to the Google Groups 
"open-iscsi" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/open-iscsi?hl=en.

Reply via email to