how to convert subcriber to publisher in UCCX 5.0

[rehan_uet]

We had high availability UCCX cluster and now publisher has been crashed, how to convert subcriber to publisher and how early this activity has to be done to save the data.

[anchoudh]

Hi Rehan,

Navigate to UCCX Appadmin->System>DataStore Control center, here you will see your UCCX Publisher\Subscriber, need to slect the Subscriber and click on Activate publisher for each of the datastores. I think it might ask you to perform the Node Manager\ Engine restart after this operation, please do so.

Please note: I would request you to please keep the Backup (using Backup and Restore) of the UCCX system before performing this operation, in case of any unrecoverable instances you an reply on this latest valid backup.And also perform this activity in nonproduction hours if its in production.

Hope it helps.

Anand

Please rate helpful posts !!

[rehan_uet]

Hi Anand,

Thanks for your response.

As license files were obtained against the Publisher (which has been crashed now), when i 'll convert the subcriber to publisher

-     will it use the same license files automatically and I dont need to add it again

-     Can i rebuild the old publisher (which has been crashed) and add it as a subcriber in the cluster with the newly configured publisher

[anchoudh]

Hi Rehan,

In this case, you can perform Remove Node Operation , and Remove the current Publisher. But before this make the Subscriber as Publisher as stated in the above post.

Please refer the section "Removing a Server" from page 301 onwards in the below link.

http://www.cisco.com/en/US/docs/voice_ip_comm/cust_contact/contact_center/crs/express_5_0/maintenance/admin/crs501ag.pdf

And again you rebuild the system with the same build and perform Add to cluster operation to make it a HIgh Availability setup.

-     will it use the same license files automatically and I dont need to add it again

Answer: Yes, existing licenses will work.

Hope this helps, also please don't forget to rate all the helpful posts by clicking on the stars below the posts.

Thanks,

Anand

[rehan_uet]

Here is the summary of steps that i'll have to go through; please confirm so that i can plan this activity

-     Backup the subcriber through Back up & Restore utility

-     Activate the subcriber to Publisher by using the following procedure

     Navigate to UCCX Appadmin->System>DataStore Control center, here I will see UCCX Publisher\Subscriber, need to      slect the  Subscriber and click on Activate publisher for each of the datastores. I may have to restart the Node Manager\      Engine after this operation

-    When subcriber is converted into publisher it will automatically start using the license file already loaded into old      publisher and there is not need to add it again.

-    If data is lost during this operation the restore the data from backup file

-     Once the new publisher is active; i'll remove the old publisher as per the procedure mentioned on page 301 of admin      guide

-     Rebuild the old publisher and add it as a subcriber in the cluster; i believe i dont need to perform any backup restore      function on it; it will automatically sync  with the new publisher

[anchoudh]

Hi Rehan,

Yes, looks fine.

One point to note here is, if you take the Backup of the UCCX HA setup it will take the backup of the entire HA setup (you can not take backup of the individual UCCX nodes in HA deployment). Inorder to Restore the backup file you need to have the same setup as it was when taking the Backup.

For your last point, you have to perform the Add to cluster operation one more time, you don't need to perfrom Backup and Restore here. Once You perfrom the Remove Node, it will become a single Node and after performing add to cluster operation one more time it becomes HA setup again, you will see that change in the Node ID's. Currently you have 0 and 1 Node ID's after the above procedure it will be 1 and 2.

Hope it helps.

Anand

Please rate all the helpful posts !!

[rehan_uet]

a little confusion, please clarify it

to restore the backup; do i need to have the same setup as it was before, you mean old publisher (which has been crashed) should be again with publisher role and this new pubhlisher will go back into publisher role.

Or

Once the subcriber is converted into publisher and old publisher is removerd and then added as a new subcriber; now we have a HA setup again; can i restore the data with this setup.

[anchoudh]

Hi Rehan,

For Restoring the back up data you need to have the same setup as before. Publisher and Subscriber needs to be same as while taking the backup.

In your case, you to be at the safer side, you are keeping a backup of the system at a valid state.

You can do one more thing.

1. Take a backup with the current setup.

2. Take one more backup , after Remove node operation is done successfully.

3. One the New HA setup is built start taking new backups.

Note: If your Remove Node and Add to cluster (after rebuilding the new system with the same Build) operations are successful, you don't need any backup and Restore operation at all.

At each stage, you need to have the same setup to Restore it.

Hope it helps.

Please don't forget to rate all the helpful posts by clicking on the stars below the right answers.

Thanks,

Anand

[anchoudh]

Hi Rehan,

Were you able to perform this operation successfully?

Thanks,

Anand

[rehan_uet]

Hi Anand,

Thanks for your follow up; before this activity something very strange has been observed. Initially when i was planning to rebuild the publisher the CRS Node Manager Service was not coming up and server was keep on rebooting with error message that "CRS Node Manager has experience an un unexpected error and system is going to reboot". As you know the CRS Node Manager is by default set to automactic and comes up automatically with system restart and was having above mentioned error message every time and system was keep on rebooting. Then i planned to rebuild this server.

Now what i did is I changed the default settings of CRS Node Manager service from automatic to manual and tried to start the service manually and it came up but what happening is after every three to four hours all the agents logout and they login again. In the control center i could see that Master roll of these server is changing in every three to four hours and during this swap of roll agents logout. One time i also observed that now original subcriber has the same issue as i had with publisher that it is keep on re-booting. Now i have shut down the publisher and subcriber is going fine.

My understanding is both servers are out of sync and creat problem after some time when both are in production. when one server is in production at a time then there is now issue.

The only issue which i have seen in the 'datastore control center' is that under 'agent' data store i see the message

"Error occurred while performing the operation. The cluster information and subscriber configuration does not match. The subscriber might be dropped (Please check SQL server log for more details).."

Rest of the datastores are OK. I tried to deactivate the datastores to re-activate them to rectify the possible database replication issue but wnen i uncheck the agent, repository and historical datastores on component activation page and click update button the next screen appears "please wait while the results are retrieved" and gets stuck there and datastores are not deactivated.

I think by re-activating the datastores may resolve the synchronization issue, but datastores are not deactivating; is there any way to synchronize the both servers or any other suggestion to resolve this situation. 

[anchoudh]

Hi Rehan,

Give a try with Re-activating the Datastores, if not please open a case with Cisco TAC, and get the entire procedure (removing the node and adding it back) done from them.

Thanks,

Anand

[rehan_uet]

Hi Anand;

I tried to re-activate the stores; after (unchecking agent, repository & historical datastores and clicking 'update') about one hour later only historical & repository datasotores were de-activated. I againt tried twice to de-activate the agent datastore but no luck even waited for hours. After that i tried to activate the repository & historical datastores but now these datastores are not activating. This all activity was done with all agents logged out and no other activity on the system. Now I have stopped all services on subcriber and publisher is working fine as a stand alone server.

This system is hosted by one of our customer in our premises and unfortunately does have a valid contract with cisco to open a case. Any idea how to proceed to get out of this situation.

Thanks,

Rehan

[anchoudh]

Hi Rehan,

You can try the below workaround, it might work.

Open   the UCCX servicability utility (from start->Programmes->Cisco   Unified CCX Administrator->Cisco Unified CCX Servicebility Utility).

Click on the Process Activation tab.

And   see which is the service which is bouncing\faling by checking each of   the services from the drop down list and check for the failure count if   it is more than 0 than it means it has failed to comeup, and if it is  in  disabled state click on Enable option there. Like wise check it for  all  the services andenable those ones whihc are falied to come up. In your case its the UCCX second Nodes Datastores, please check each of the Datastores for the subscriber and activate it.

And you need to Restart the UCCX Node Manager to take this effect.

Than check the behaviour.

Hope it helps.

Anand

[anchoudh]

Hi Rehan,

Any update on this issue? Were you able to come out of it?

Thanks,

Anand