On 03/02/12 08:41, Ulrich Windl wrote:
> [...]
>   
>> > Users have only been hitting it during shutdown at reboot time though. I
>> > think what was happening was the initiator sent a logout, got a logout
>> > response, but then the target still sent data/responses for other IO
>> > that was running at the time. The initiator assumed that once the logout
>> > response was sent the target would stop sending data.
>>     
> [...]
>
> According to my experience this can happen with multipath: When you use 
> "multipath -f map", the command returns quickly, but actually patch checkers 
> are removed in background after that. So if you disconnect your iSCSI 
> immediately after flushing your maps, you may hit this. I haven't had this 
> with iSCSI, but with a regular SAN (where I also had hard hangs in I/O).
>   

Thanks for your responses. I'll try this. Luckily, this hasn't too high
impact for us.

Unfortunately, we have AUFS on our disk-less servers. After rebooting
everything is gone. So the display from IPMI is the only thing I have at
the moment. We need to change this.

Cheers,

Sebastian

-- 
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