CUCM Upgrade procedure question

Unanswered Question
Feb 27th, 2009

I`m due to upgrade a CUCM 6.X server to 6.1. and I`m querying a procedure

Current set up

Publisher - Running as Glass House- No CM Services, No TFTP etc

Subscriber

2x TFTP Servers

I know we upgrade to an inactive partition so not to effect production however due to various reasons I would like to do the following

Pull Disks on Pub to ensure we have a physical roll back option- test each disk in turn

Change Pub to allow the test phones to register to it

Upgrade Pub on Inactive Partition

Allow test phones to register to Pub - make test calls, access Admin , EM testing etc

Question

As the phones will reboot and so query the TFTP`s which I have not upgraded to the newer SW, it will then register to the Pub which is now on the newer SW- will this work or will it be a conflict between the Pub being on the newer SW and the TFTP being on the older version ?

or should I manually reconfigure the test phones only to use one of the TFTP Servers and the upgrade the same TFTP after I`ve done the Pub and reboot both- PUB and TFTP on the newer SW

Once I`m happy we will then upgrade the other TFTP followed by the Subs

Any thoughts or opinions

thanks

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
Jaime Valencia Fri, 02/27/2009 - 04:08

If by this:

Change Pub to allow the test phones to register to it

Upgrade Pub on Inactive Partition

you mean switch version on PUB and then upgrade SUBs (no switch version, but install upgrade) that's wrong

this applies to any linux upgrade:

Upgrading from Unified CM 5.x to Unified CM 6.x

http://www.cisco.com/en/US/docs/voice_ip_comm/cucm/srnd/6x/callpros.html#wp1043707

HTH

java

if this helps, please rate

iptuser55 Fri, 02/27/2009 - 04:49

Java

I mean to use the Pub as a quick test and if successful then upgrade and switch on the remaining Subs.

If the upgrade test is not good then re-insert the disks and reboot Pub /TFTP or just switch back on the Pub

Jaime Valencia Fri, 02/27/2009 - 05:03

the thing is that even if it's successful you'll need to switch back before upgrading the SUBs, if you leave PUB on 6.1(2) for testing and the SUBs in 6.1(1) and then try to upgrade you won't be able to do so

that's why the instructions clearly say to install the upgrade in all servers at a time, then switch one at a time

HTH

java

if this helps, please rate

iptuser55 Fri, 02/27/2009 - 06:03

Java

the reason I`m asking if that we had a failure when reverting back to the lower SW in our test lab so the idea is to the test the roll back on the Pub only first in case the lab and production Cluster are any way differetn though they are on the SW level. Would not the switch back to the Pub to the lower SW so the cluster is now all on the same Verison and then upgrade the Subs- once all done then switch one by one as per doc?- would this not work

Actions

This Discussion