CSCvg76031 - IMP services stuck in starting state after fresh install/upgrade

[Michael Stark]

I am doing a PCD upgrade for the IMP servers and it cannot continue after the first server because the (PAWS) service is down amongst a few others on the primary.  Using the 12.5(1) latest upgrade file.

What is the fix for this Cisco?

[Thilina]

Hey mate , thanks.

Yeah , I already tried that. but no luck :(

[Joe Christensen]

Ended up having to open a TAC case. It was an issue where the server was built with the hostname capital letter, node name in lower case, and it caused issue with the hosts file. TAC had to get root access and issue a command to modify the host file. 

[Thilina]

Thanks mate , I will open a TAC case as well then. It could be the case for me as well.

Regards

[GursimratPalSinghKhaira3002]

I know this is OLD thread and probably might have solved this issue just for any other user how faces this issue for Cisco Presence engine stuck in starting

Presence Engine Service stuck in STARTING state after 12.5 fresh install or PCD Migration

CSCvg94247

[Wes Sisk]

Beyond the referenced bugs many times this symptom of service stuck starting is related to DNS and name resolution. From the admin CLI use the command "show tech network hosts" and pay special attention to the /etc/hosts and localhost sections. Any variations here will likely require a TAC case and remote access to resolve.

Things to check without TAC:

1. ensure forward and reverse DNS resolution of the server name, ensure server is pointing to correct DNS server. Fix any DNS issues and reboot the servers.

2. examine the output of admin CLI command"show tech network hosts".

If you discover any inconsistencies then open a TAC case for further assistance.

[Dalton-Covene]

Great post Michael.  +5

I ran into this scenario when upgrading from 10.5.2 to 12.5.1.  After issuing this command on each IM&P node, the services started within 5 minutes. No reboot required.

[Michael Stark]

Great to hear this worked for you too :)

[Gerry O'Rourke]

Be careful with this command.

While disabling the replication-sync will allow the services to start up - I do not believe it is not the root cause of the issue.

The reason the services are not start is likely because the database replication is not started or completing.

When replication-sync is enabled it is designed to not allow other services to start:

"The Cisco Replication Watcher service blocks other services from starting until database replication is setup and functioning normally."

CLI Reference 

The issue I experienced (12.5 SU4) was that the subscriber IMP server did not ask for replication to be initialized.

To fix the issue I had to reset dereplication.

I recommend you check using utils dbreplication runtimestate

And use this reference if necessary to fix.

Gerry