We have 4 XenServer 6.0 servers connecting via iSCSI to our EMC VNX5300 SAN. We are constantly getting Multipath Alerts in the XenCentre console reporting that there has been a change in multipath status.
When I check the SP logs on the SAN I can see that the iSCSI Initiators are performing a logout and then a login about 5 seconds later. The logout has an event id of 0x7117000e and then the login event id is 0x7117000d.
Although it doesn't look like this is causing a major issue I would like to correct the configuration so that the multipath settles down.
Has anyone seen this sort of behaviour?
The errors 0x7117000e/d are just generic warning for login/logout:
iSCSI Login - Initiator Data: <initiator data> Target Data: <target data>
iSCSI Logout - Initiator Data: <initiator data> Target Data: <target data>
For troubleshooting we need more details of the error. Check in the logs SP event, SPcollect or Ktrace with the same timestamp, and put it here for us.
Please see Support Solution emc156408 on PowerLink for Best Practices for connecting the array to the hosts when using iSCSI. Also see emc245445 that lists all the articles about iSCSI and issues that you may encounter.
When you see a lot of log in and out messages, that can be a couple of things - network configuration is the most likely - how the hosts and NIC's on the hosts are configured and the use of different subnets for each NIC and SP ports pairs (a port on SPA and SPB can share the same subnet, but that the same subnet as other pairs).
I am assuming that your iSCSI IO modules are resetting with error code 0x7117000b.
In any case, please check if all the initiators are on same subnet, if thats the case then you need to change it according to EMC best practices using the support solution provided by Glen above .
I had something similar happening in my environment, we did the Re-IP and everything seems fine now.