On 06/30/2012 09:07 PM, Mike Christie wrote:
On 06/30/2012 12:53 PM, Nik Martin wrote:
Linux 2.6.18-308.8.2.el5xen #1 SMP Tue Jun 12 10:39:28 EDT 2012 x86_64
x86_64 x86_64 GNU/Linux
The initiators have three ethernet ports active, management, public, and
storage, all intel GBEs, Connected to the same Dell Powerconnect.
So are you running the initiators in rhel/centos 5? What version? If you
run them in the newest version or run them in 6.2 does it work ok?
If you are using a older version of rhel5 then you might be hitting a
bug where we mis detected a bad network. The initiator would not see IO
being processed and would send a iscsi nop as a ping. The ping would
fail to execute within your noop timeout settings. But it would fail
because you were pushing a lot of IO and the nop/ping just got stuck. We
fixed this by checking if other IO sent before it was completing then we
gave the nop/ping more time.
Mike,
I am running the initiators on CentOS 5.8. I downloaded and built the
latest stable version and once I figured out why it wasn't logging in
automatically (thanks), it seems to do OK, so this maybe the issue.
Small, bursty transactions were ok, but moderate to heavy loads would
wreak havoc. I am also suspicious of my network config, so this may be
a red herring. The 10G CX4 modules from Intel are new to me, and I'm not
100% sure a stock CentOS is handling the network flow control on the
target very well. I have several errors on teh switch port they are
connected to.
--
Regards,
Nik
--
You received this message because you are subscribed to the Google Groups
"open-iscsi" group.
To post to this group, send email to open-iscsi@googlegroups.com.
To unsubscribe from this group, send email to
open-iscsi+unsubscr...@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/open-iscsi?hl=en.
