We have a distributed unity 10.5 cluster (partner installed) where the subscriber seems to have failed and I have engaged Cisco TAC for help. We are now 6+ weeks in trying to get a new subscriber to come up and be an active participant in the cluster. This cluster supports about 600 phones and users tied together over a spoke and hub type of MPLS structure covering approximately 25 spokes and 2 hubs with a SIP trunk coming into each hubs. The cucm and unity clusters have a subscriber or publisher at each hub.
The publisher and subscriber hosts were initially side by side at one hub. As part of the final acceptance of the 500k project the subscriber was moved to another location. It seemed to run but about 2 months ago I noticed that the GUI was non-functional on the subscriber and opened a TAC ticket.
Can anyone respond to the following
(1) The primary and secondary servers in a cluster must match in CPU and amount of memory, or the configuration is not supported. Does this mean that the CPU chips supporting the two VMware hosts must be the same? We have a publisher based on a E5-2620 2ghz 6 core processor running in a UCS C220-M3S box and the subscriber is based on a E5-2609 2.4Ghz 4 core chip running in a UCS C220-M3SBE box
(2) Can a cluster be built where the subscriber is on a vmware 5.1 host and the publisher is on a 5.5 host? I know there is talk that in a 5.1 box you must leave one vcpu available for Vmware scheduling, but that this is not the case for a 5.5 host. On the two hosts there is one non-committed vCpu for host.
(3) Does the following bring anything to mind based upon peoples past experiences.
after rebuilding the subscriber with it defined as a member of the cluster in Unity GUI, I issue the following on both cluster members and get the responses listed below
command issued: "show cuc cluster status"
publisher
ACE_File_Lock::ACE_File_Lock: Permission denied /dev/shm/CCM_GENstatusLock_0
Server Name Member ID Server State Internal State Reason
----------- --------- ------------ ----------------- ------
unity 0 Primary Pri Single Server Normal
Database replication is not active
admin:show cuc cluster status
Server Name Member ID Server State Internal State Reason
----------- --------- ------------ ----------------- ------
unity 0 Primary Pri Single Server Normal
Database replication is not active
Subscriber
admin:show cuc cluster status
Incorrect password or user com.informix.asf.IfxASFRemoteException: cucli@tempeunity is not known on the database server.
Command failed
admin:show cuc cluster status
Incorrect password or user com.informix.asf.IfxASFRemoteException: cucli@tempeunity is not known on the database server.
Command failed
(4) Since the subscriber was configured to be part of the cluster from initial configuration, are there some other types of health related CLI commands that can be run to determine why a successful install has failed to produce a working cluster?
