CUCM Publisher data restoration in a cluster

[hvDinoBot]

I have a Publisher and two subscribers cluster.

CUCM Pub failed. I ended up rebuilding v11.5 CUCM Publisher and restored the data from a backup. This was done twice with the same results. The Publisher DB is not in sync with the subscribers. Missing a step, or should I use another method?

On Publisher, it's showing the below.

Cluster Detailed View from pub (3 Servers):
PING DB/RPC/ REPL. Replication REPLICATION SETUP

SERVER-NAME IP ADDRESS (msec) DbMon? QUEUE Group ID (RTMT) & Details
Pub   20.13.0.15 0.014 Y/Y/Y -- (-) (-) Replication Not Setup
Sub1 20.13.0.16 0.070 Y/Y/Y -- (-) (-) Replication Not Setup
Sub2 30.13.4.16 0.299 Y/Y/Y -- (-) (-) Replication Not Setup

On both subscribers, it's showing the below

pub 0 0 - initializing
sub1 706 2 - good
sub2 706 2 - good

Steps I have done to restore - MISSING a step?
Stop DB replication on both subscribers.
Rebuild v11.5 CUCM Publisher with all the same credentials
Once Pub is completely UP, map the SFTP server and start the data restore.
Pub data restored, then reboot it first until it's completely up via CLI
Reboot Sub 1 and wait until it's up
Reboot Sub 2 and wait until it's up.

[Roger Kallberg · Accepted solution]

Very last step in the DRS admin guide for First node restore has this note.

To see more details on DRS restore please see this document. Disaster Recovery System Administration Guide for Cisco Unified Communications Manager and IM & Presence Service, Release 10.x 

[hvDinoBot]

I have been eyeing that command for the past couple of days. Is this only runnable during off hours, or can it be run during production time and, of course, with no one touching/making changes in the CUCM servers? I will like to do it soon. I also mentioned that I did run the util dbreplication repair all command the day before, but that didn't make a change. The one you mention I will need to do

[hvDinoBot]

After utils dbreplication rebuild all, I used utils service restart A Cisco DB then started to see the results from Replication Not Setup
to Setup Completed
Thank you,

[Roger Kallberg]

Glad to hear that.

[mradell]

To add to what Roger posted, you need to restart the nodes (Publisher followed by Subs) per the Disaster Recovery Admin Guide. I'd do that first unless you already have, then force the replication if needed:

per https://www.cisco.com/c/en/us/td/docs/voice_ip_comm/cucm/admin/12_5_1/admin/cucm_b_administration-guide-1251/cucm_b_administration-guide-1251_chapter_010100.html#CUCM_TK_RD5B402C_00

[hvDinoBot]

The Pub and two subs were rebooted after the restore process. I might have to do it again. Is the utils dbreplication build command possible during business hours or not a good idea? 

[mradell]

As far as I'm aware it should not affect production. That said, waiting until a downtime to cover yourself is always best unless it's creating an issue you need to fix immediately.