UCM down/features disabled CUCM 9.1/7945s

[drewdown]

Starting to see these messages popping up since we migrated to a single building from two. We have a cluster and when it first showed up after moving I/O wait was high on the PUB so we rebooted it and all phones registered to the SUB but the problem did not go away. ARP timeouts are default 4 hours, no network changes, CUCM is running on vmware and no changes there either that I am aware of. Whe we moved buildings we simply moved all the users to 1 and left the core(dual 6500s)/demarcs in the other. So if anything the load on the cross link l (4 x 10G) increased but not by all that much (200 users total).

From my limited knowledge it seems that the phones are losing connectivity to CM (both pub and sub) but no drops or timeouts that I can see. It then leads me to believe it may be vmware related? One of the network engineers was on the phone with Cisco when the problem happened to him (lol). Anyways could it be backend storage? Anyone seen this and have it not be path or route problem?

 

[gnrslash4life]

It sounds like keep alives are dropping which would point to some network related issue. Also, you want your phones registering to the sub not the pub.

Certifications:
A+
Network+
CCENT
CCNA Voice
TVOICE
CAPPS
CIPTV1
CIPTV2

 

[drewdown]

Ok so they are registered to the SUB so that is ok? I thought PUB was primary, why register to the SUB? At this point the only thing that changed was rebooting the PUB and consolidating the buildings, so ifs not either of those no idea what it could be. Looking around the internet it seems like people who do have this problem have a hell of a time tracking it down.

 

[gnrslash4life]

The PUB is the primary that pushes all the updates and what not. You never want devices registered to it. If I had to guess, I would say you have an issue with your connection connecting the 2 buildings.

Certifications:
A+
Network+
CCENT
CCNA Voice
TVOICE
CAPPS
CIPTV1
CIPTV2