Unsolved
This post is more than 5 years old
2 Posts
0
67222
PS6000 iSCSI SAN connectivity problem with Vmware ESX hosts
This query is in relation to a PS6000 iSCSI SAN we now have in production with R900's. We have the iSCSI traffic on a completely separate physical network. We have ports from the 2 x R900's and the PS6000 on this ‘iSCSI network'....so we have the ESXi kernels connected to the iSCSI network and 3 (active) ports on PS6000 connected....We have a dedicated management port on the PS6000 on our standard production network.This morning there was broadcast storm on our standard production LAN....but for some reason this seemed to effect the traffic from one of our Hosts to the PS6000 on the iSCSI network...the Host lost connection to 3 volumes and we just can't figure out why.Below is the error's that kept appearing in the event log of the SAN, these continuted until the boradcast storm ended the host then re-established communication with the Hosts. As stated the other Host has the same configuration and was unaffected.We cant see any reason why the ISCSI network was affecteded as its on its own physical network (we have checked the cableing on both boxes and verified this). The management port on the SAN is restricted to management acces as well. So we dont see how this occured and why it only happened one of the boxes.The error we seen in the SAN logs are below
-----------------------------------------
ERROR event from storage array PS6000DUB01
subsystem: MgmtExec
event: 7.4.3
time: Wed Sep 30 07:25:25 2009
iSCSI login to target '10.10.10.22:3260, iqn.2001-05.com.equallogic:0-8a0906-0b703fd04-f400000035d4aa4e-dubsangrp01vol02' from initiator '10.10.10.31:60880, iqn.1998-01.com.vmware:dubvirsrv01-3e8012bb' failed for the following reason:
Connection login timed out
-----------------------------------------
ERROR event from storage array PS6000DUB01
subsystem: MgmtExec
event: 7.4.3
time: Wed Sep 30 07:25:37 2009
iSCSI login to target '10.10.10.21:3260, iqn.2001-05.com.equallogic:0-8a0906-2b603fd04-7710000038f4aa4f-dubsangrp01vol01' from initiator '10.10.10.31:60888, iqn.1998-01.com.vmware:dubvirsrv01-3e8012bb' failed for the following reason:
Connection login timed out
-----------------------------------------
ergosteur
2 Posts
0
November 26th, 2009 10:00
I have EXACTLY the same problem.
4x PowerEdge R710 hosts
1x PS6000 EqualLogic iSCSI SAN
Rebooting a host sometimes causes the LUNs to show up, and sometimes causes other ones to disappear. Currently we have one host fully up, one that's missing one LUN, and the other two can't see the SAN at all. I've tried removing the iSCSI initiator, changing the WWN, removing and recreating the ACLs on the SAN, nothing works.
Did you manage to resolve the issue?
janardhanr1
1 Message
0
July 19th, 2010 23:00
Did any 1 solve this? All my luns are not visible now?
Any help is appreciated.
Thanx
Jana
ergosteur
2 Posts
0
July 20th, 2010 08:00
Hi,
In our case, the problem was solved by connecting via the serial console and doing a couple of proper shutdown/reboot cycles. It seems the connections to the hosts were not closed properly, and this forced the SAN to gracefully kill all connections.
Anonymous
5 Practitioner
5 Practitioner
•
274.2K Posts
0
August 24th, 2010 10:00
Hmmmm... That's not sounding right. Killing connections don't require a shutdown/restart.
What version of firmware are you using? For PS6xxx series it should be 4.3.6.
-don
ianroche80
2 Posts
0
August 25th, 2010 09:00
I actually got this sorted the box was fully patched when this occured and it had the fix you mention. The problem was one of the LUNs had been put offline but the host still had a path to it showing as DEAD. I done a tail -f /var/log/messages on the host console when an outage occured and it was full with messages of the host trying to connect to this lun. As soon as I removed this path the issue resolved itself.
mfkelly_tx
31 Posts
1
August 25th, 2010 09:00
There was a patch that addressed this, we had the problem at one time, but have not seen it for quite some time as I try to keep vSphere hosts/vCenter up to date.
http://communities.vmware.com/message/1507794#1507794
http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1019497
In a software iSCSI environment, the iSCSI connection might go offline when multiple VMkernel NICs in the same subnet are used to access a target.