Warning message : Ungraceful shutdown detected ( cucm/unity/pressence)

[iptsupport]

Hey Everyone,

i have a customer with 2 ucs220 server in sperate side that we installed for him with CUCM/CUP/CUC ( all version 12.5su4)

in one of the side the servers shutdown because power problem in the building and after that all the servers in the side ( all the servers in the side is Subscribe) up with this warning message :

WARNING: Ungraceful shutdown detected - A rebuild of this node is highly recommended
to ensure no negative impact(such as configuration or file system corruption). For
rebuild instructions, see the installation guide.

the meaning of that is to install the node from the beginning?

there is any other way to rebuild the node without install new one?

[Nithin Eluvathingal]

To remove  the warning rebuild is needed.. its an enhancement in SU(4). Have a look on this Bug Id CSCvy68211

[iptsupport]

Hey Nithin,

i know that someone is open TAC to Cisco about that but i thought maybe someone is find way to solved that without rebuilding the server.

Did you know what is necessary before rebuild CUCM/CUC Subscribe node? ( any decontamination or video can help)

for IMP i found video but i didn't find any thing about rebuild CUCM/CUC Subscribe node.

[Nithin Eluvathingal]

A rebuilding  redeploying a new VM with restoring the DRF backups. Refer below  guide https://www.cisco.com/c/en/us/support/docs/unified-communications/unified-communications-manager-callmanager/214626-configure-backup-and-restore-from-gui-in.html 

[iptsupport]

Hey Nithin,

are you sure that rebuild node is when you redeploying a new VM with resorting backup?

because i did the processes of redeploying VM with restoring and im not delete like cisco the presence from the CUCM Publisher GUI.

i attach the video of cisco : https://www.youtube.com/watch?v=yUomIrSUW-A 

[Roger Kallberg]

What do you mean by this “im not delete like cisco the presence from the CUCM Publisher GUI.”?

[iptsupport]

Hey Roger,

in the video of Cisco if i understand that they show process that you need to do when you rebuild  your subscriber.

i attach the video like you can see.

in the past i have customer that is UCS machine stop working so we need to install  all the applications in another UCS after we install them we did restorer from backup that we have without do all the prepossess that Cisco show in the video.

so I'm not think that when Cisco say rebuild they mean to  re-install the machine and after that restore from backup.

[Roger Kallberg]

Sorry but it is very hard to understand what you write.

A rebuilt or restore is pretty much the same thing in my book and it includes installation of a new VM, with the same name and any other information like IP address, security password and so on…, and then a restore from DRF, aka restore a backup of the node.

[Nithin Eluvathingal]

Rebuilding  means  reinstalling  the server on new VM and resorting the BACKUP. 

You can try  @Kathy N.  option by upgrading the cluster to 12.5 SU(5). 

[Gilberto Millan]

Hi @Kathy N. ,  Hi @Nithin Eluvathingal , if I take the option of the update, my only doubt is what about the Tac support?

I have read on another post that the cluster could be out of support, what was your experience with this ?

[Nithin Eluvathingal]

Two weeks back I had a similar situation with a customer site and I did a reinstall from the BACKUP. 

I never opted a  TAC ticket for this  issue. If you have valid smartnet contract open a ticket with Cisco TAC and  take their suggestion. 

[Adam Pawlowski]

The rebuild process typically is to re-deploy the VM, or mount the install media and re-install the appliance. After re-installing it, a DRF restore can be used to return it to the cluster. It seems acceptable to use a backup to restore the appliance, in that I guess the contents of the backup are not deemed to be subject to the same data corruption concerns as the operating system and software itself.

If you do not have a DRF backup, and are rebuilding and re configuring the appliance from scratch, then you'd follow whatever instructions apply to removing/adding nodes. For IM & Presence, there's a call out on removing the node explicitly before attempting to re-add it fresh, due to some sort of keys exchanged between the two things which will not allow you to simply install it fresh and have it join the cluster.

I believe there's an enhancement as well for some sort of file system consistency check, but this doesn't exist today.

If you have a backup it is really not very difficult, just some hours of time.

[iptsupport]

Hey Adam,

first i want to thank for you answer, this very helpful for me.

about DRF Backup, the schedule backup that we configure stop working before month ago so we afraid to using that.

about rebuild the server, i try to search any decontamination about rebuilding CUCM Subscribe or Cisco unity Subscribe and I'm not found anything.

about IMP i found this video that explain exotically how to do that : https://www.youtube.com/watch?v=yUomIrSUW-A

did you see and decontamination about rebuilding new subscribe node of CUCM/CUC?

did you now what is necessary to do before rebuild new subscribe node?

[Adam Pawlowski]

I believe you can still take a current DRF back up and restore from it. When I'd asked TAC about my own scenario, they were not concerned about the data in the backup being problematic.

I believe for Unity Connection and the UCM you can delete and rebuild the subscriber guest as though you were installing it new. That would mean platform specific configuration would need to be done again, non replicating files copied over, etc. But the cluster database and such should replicate over. 

If you can take a new DRF backup now, then re-install and restore following recovery procedures, I believe this is fine. This is what I've done. 

The video you posted on IMP shows the process to rebuild the subscriber, as opposed to restore. If you don't have a backup and cannot get one, then this would be the option. For IMP I believe you can export contact lists and such as well if you cannot get a DRF backup, which can be re-imported, but obviously it would be preferred to have a system backup.

[Kathy N.]

In addition to a server rebuild, an upgrade will also remove the warning message.  If you've been postponing an upgrade or have one available, this may be an easier solution than rebuilding the servers.  We did an upgrade but shortly afterwards had another unexpected power outage so we've learned to live with the warning message.  Hopefully one of the next releases will provide a way to acknowledge this alert to clear it.