On 03/30/2011 05:20 AM, Matthew Richardson wrote:
On 30/03/11 05:21, Mike Christie wrote:
On 03/29/2011 06:51 AM, Matthew Richardson wrote:
- Is it possible to turn on the libiscsi2 and iscsi_tcp debugging at the
same time?
Done - log attached
11:03:20 unplug
11:03:30 replug
11:09:07 disk unblocks
- Could you also take a tcpdump trace on the target, so we can check
that the target is getting the scsi commands that we are waiting on?
pcap file attached - (140.10 is target, 140.3 is initiator)
It looks like we are overqueueing after we relogin (we relogin to handle
the cable pull). It looks like the overqueue could be due to how we
reset the values after we relogin. We reset some of the values when we
should not and that causes us to think the target's window is larger
than it is. The target then drops the excess commands. On our side the
commands then timeout and the scsi eh runs and we see the second conn
error as a result of that.
I cannot replicate this here. The window seems to stay open. What is
your QueuedCommands value in /etc/ietd/ietd.conf?
If I send you a patch can you build a kernel to test? If not I can build
a rhel based rpm with the patch in it for you to test. I think you can
use rhel rpms with Scientific Linux, right?
--
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.
