zfs discuss - Feb 2008 - iscsi core dumps when under IO

Hello,
 I am getting a strange issue when using zfs/iscsi shares out of it.

when I have attached a a cent os 5 initiator to the zfs target it works fine
normally until i start doing heavy 100MB/s+ copies to a seperate cfs/nfs export
on the same zfs pool.

the error I am getting is:

[ Feb 20 10:41:07 Stopping because process dumped core. ]
[ Feb 20 10:41:07 Executing stop method ("/lib/svc/method/svc-iscsitgt stop
143") ]

I was wondering if any one had any ideas. I am running 10U4 with all of the
latest and greatest patches. Thank you.
 
 
This message posted from opensolaris.org

Stephen,
> I am getting a strange issue when using zfs/iscsi shares out of it.
>
> when I have attached a a cent os 5 initiator to the zfs target it  
> works fine normally until i start doing heavy 100MB/s+ copies to a  
> seperate cfs/nfs export on the same zfs pool.
>
> the error I am getting is:
>
> [ Feb 20 10:41:07 Stopping because process dumped core. ]
> [ Feb 20 10:41:07 Executing stop method ("/lib/svc/method/svc- 
> iscsitgt stop 143") ]
>
> I was wondering if any one had any ideas. I am running 10U4 with all  
> of the latest and greatest patches. Thank you.
There are a set of issues we have recently have been resolved in  
Nevada regarding the iSCSI Target under load. We are looking at back  
porting these changes to S10.

The nature of the failure appears to be an iSCSI Initiator seeing long  
service times (in seconds), triggering a LUN reset. The LUN reset  
causes all I/O to be cleaned up specific to that LUN. Given the multi- 
threaded nature of the iSCSI Target, the odds are pretty high that  
cleanup across every possible I/O state would be possible, and some of  
the states were not handled correctly.

The follow command is likely to show the reason for the "process  
dumped core", being an assert in the T10 state machine.

# mdb /core
::status
::quit

> This message posted from opensolaris.org
> _______________________________________________
> zfs-discuss mailing list
> zfs-discuss at opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Jim Dunham
Storage Platform Software Group
Sun Microsystems, Inc.

http://blogs.sun.com/avs
http://www.opensolaris.org/os/project/avs/
http://www.opensolaris.org/os/project/iscsitgt/
http://www.opensolaris.org/os/community/storage/

I tlooks like you are probalby right.

from MDB this is what i''m getting:

file: /usr/sbin/amd64/iscsitgtd
initial argv: /usr/sbin/iscsitgtd
threading model: multi-threaded
status: process terminated by SIGABRT (Abort)
panic message: Assertion failed: 0, file ../t10_sam.c, line 511

Would it be possibly fixed in the S10U5 release?
 
 
This message posted from opensolaris.org