ietd died but connections are still up.

[Emmanuel Florac]

Hi, I had a strange problem today : when restarting 2 (windows)
initiators, they couldn't reconnect to the target (0.4.16 on 2.6.22.19,
am64), though 4 other initiators were still doing fine.

Actually the main ietd daemon died, though the connections were still
up. Nothing appears in the logs, except quite a lot of iscsi lun reset
(exactly 75 actually) that occured for no apparent reason (no hardware
problem):

<6>iscsi_trgt: Logical Unit Reset (05) issued on tid:2 lun:2 by
sid:7882399782338624 (Function Complete) <6>iscsi_trgt: Logical Unit
Reset (05) issued on tid:2 lun:2 by sid:7882399782338624 (Function
Complete) <6>iscsi_trgt: Logical Unit Reset (05) issued on tid:2 lun:1
by sid:7882399782338624 (Function Complete) <6>iscsi_trgt: Logical Unit
Reset (05) issued on tid:2 lun:2 by sid:7882399782338624 (Function
Complete) <6>iscsi_trgt: Logical Unit Reset (05) issued on tid:2 lun:1
by sid:7882399782338624 (Function Complete) <6>iscsi_trgt: Logical Unit
Reset (05) issued on tid:2 lun:0 by sid:7882399782338624 (Function
Complete) <6>iscsi_trgt: Logical Unit Reset (05) issued on tid:2 lun:1
by sid:7882399782338624 (Function Complete) <6>iscsi_trgt: Logical Unit
Reset (05) issued on tid:2 lun:2 by sid:7882399782338624 (Function
Complete) <6>iscsi_trgt: Logical Unit Reset (05) issued on tid:2 lun:0
by sid:7882399782338624 (Function Complete) <6>iscsi_trgt: Logical Unit
Reset (05) issued on tid:2 lun:1 by sid:7882399782338624 (Function
Complete) <6>iscsi_trgt: Logical Unit Reset (05) issued on tid:2 lun:2
by sid:7882399782338624 (Function Complete) <6>iscsi_trgt: Logical Unit
Reset (05) issued on tid:2 lun:0 by sid:7882399782338624 (Function
Complete) <6>iscsi_trgt: Logical Unit Reset (05) issued on tid:2 lun:2
by sid:7882399782338624 (Function Complete) <6>iscsi_trgt: Logical Unit
Reset (05) issued on tid:2 lun:0 by sid:7882399782338624 (Function
Complete) <6>iscsi_trgt: Logical Unit Reset (05) issued on tid:2 lun:2
by sid:7882399782338624 (Function Complete) <6>iscsi_trgt: Logical Unit
Reset (05) issued on tid:2 lun:0 by sid:7882399782338624 (Function
Complete) <6>iscsi_trgt: Logical Unit Reset (05) issued on tid:2 lun:2
by sid:7882399782338624 (Function Complete) <6>iscsi_trgt: Logical Unit
Reset (05) issued on tid:2 lun:0 by sid:7882399782338624 (Function
Complete) <6>iscsi_trgt: Logical Unit Reset (05) issued on tid:2 lun:2
by sid:7882399782338624 (Function Complete)

Any ideas about what happened ?

--
----------------------------------------
Emmanuel Florac     |   Intellique
----------------------------------------


-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
_______________________________________________
Iscsitarget-devel mailing list
Iscsitarget-devel@...
https://lists.sourceforge.net/lists/listinfo/iscsitarget-devel

[Arne Redlich-2]

Am Donnerstag, den 19.06.2008, 12:07 +0200 schrieb Emmanuel Florac:

> Hi, I had a strange problem today : when restarting 2 (windows)
> initiators, they couldn't reconnect to the target (0.4.16 on 2.6.22.19,
> am64), though 4 other initiators were still doing fine.
>
> Actually the main ietd daemon died, though the connections were still
> up. Nothing appears in the logs, except quite a lot of iscsi lun reset
> (exactly 75 actually) that occured for no apparent reason (no hardware
> problem):
>
> <6>iscsi_trgt: Logical Unit Reset (05) issued on tid:2 lun:2 by
> sid:7882399782338624 (Function Complete)

<snip>

> Any ideas about what happened ?

No, we'd have to look at a core dump to determine why ietd died. Can you
try to reproduce this crash, possibly with gdb attached or at least with
core dumps enabled (add "ulimit -c unlimited" before starting ietd to
the init script)?

The failed reconnect is due to ietd not running, so initiators cannot
establish new sessions, while open sessions remain alive.

Thanks,
Arne


-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
_______________________________________________
Iscsitarget-devel mailing list
Iscsitarget-devel@...
https://lists.sourceforge.net/lists/listinfo/iscsitarget-devel

[Emmanuel Florac]

Le Thu, 19 Jun 2008 13:53:02 +0200 vous écriviez:

> No, we'd have to look at a core dump to determine why ietd died. Can
> you try to reproduce this crash, possibly with gdb attached or at
> least with core dumps enabled (add "ulimit -c unlimited" before
> starting ietd to the init script)?

It will be hard to reproduce : ietd crashed after working fine under
heavy load for several weeks (not that bad, the windows initiators had
to be rebooted countless times in the meanwhile :) And it's actually in
production, so I'd rather not turn debugging on but I'd try to get a
core dump.

Thanks!

--
--------------------------------------------------
Emmanuel Florac               www.intellique.com  
--------------------------------------------------

-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
_______________________________________________
Iscsitarget-devel mailing list
Iscsitarget-devel@...
https://lists.sourceforge.net/lists/listinfo/iscsitarget-devel