I already have a ticket open for this but after working with TAC for a few hours, they're still not sure what the problem is and the resolution for it. I wanted to check here also and see if someone already experienced this and got it resolved.
This happened immediately after upgrading from CUCM 11.5.1 to 14 SU3. No certificates were expired before the upgrade and of course, none are expired after the upgrade. Only one of the subscriber servers has this problem. Server has been rebooted and Tomcat certs were generated (even though they weren't expired) just to see if that would help, since other cases like this were related to expired certs. All the servers use the standard self-signed certs.
I have 5 CUCM servers in a cluster. 1 Pub and 4 Sub servers. All 5 servers were upgraded with no errors from version 11.5.1 SU11 to CUCM 14 SU3 (14.0.1.13900-155). I wanted to install the post upgrade COP file (ciscocm.postUpgradeCheck-00041.cop.sha512) to check for errors. This COP file installed with no issues on 4 of the servers, but one of the subscriber servers has an error when I go to the install/upgrade page. It won't let me install anything there. I see 'Unmapped Exception null' and the only option is to select Cancel (there's no 'Next' button). Here's the path I go to and what happens: OS Administration page > Software Upgrades > Install/Upgrade (unusually long wait, maybe over 1 minute) > Error message: 'Unmapped Exception null'
To make things more interesting (or confusing), if I go to the Publisher server and log in to the OS Admin page and then go to Software Upgrades > Install/Upgrade Cluster I am able to install this COP file in that way (I.E., using the Install/Upgrade Cluster' option from the Publisher server rather than going directly to the Subscriber server and installing it there). However, I know there are files we sometimes need to install manually/individually on each server (not using the 'upgrade cluster' option from the Pub)...so I need to get this fixed.
Have you seen anything like this yourself?...and a resolution for it?
Thanks
Accepted Solutions
Not the same but we had issue with a CUCM cluster that we upgraded to 14.Su3. Some users had missing device associations on one of the nodes, while the other nodes showed the device associations correctly. We could not associate the devices on the problematic node. TAC could not identify the root cause and suggested us to rebuild that node. That worked for us. You might want to try the same thing..
OK, just finished a call with Cisco and it seems our firewall in the location of this one Subscriber server (it's in a different location) is not able to get the SFTP server/password info from the Publisher server which is in another location and the most likely reason is the firewall is blocking it. We proved the problem Subscriber CAN do an install/upgrade from its own OS Admin webpage, but ONLY if I first go to the Publisher's OS Admin page, select 'Cluster Software Location', select the server with the problem and select the option for 'Use below download credentials and software location' INSTEAD OF 'Use download credentials and software location from Publisher'. After telling that server to use the manual entry of the software location settings (which is exactly the same as the username, SFTP server, etc that all the other servers use), it works. When you click 'Install/Upgrade' on the Subscriber's OS Admin page, it will try to contact the Publisher to get the SFTP server info and it looks like in my case, that traffic is not making it to the Publisher, so the only logical explanation is that the firewall is blocking it. Traffic TO that Sub from the Pub seems OK, but if that Sub tries to contact the Pub for this info, it's not working. Again, the only logical reason I can think of for that to happen is our firewall is blocking it, so I'll talk to the team about that.
In the meantime, we have an easy workaround.
Settings on the Publisher for this one Subscriber (temporary, until firewall is updated):
