| Author |
Message
|
| schow |
 Posted: Thu Nov 17, 2005 7:11 am Post subject: Do full repositories have to be upgraded to WMQ v6 first ? |
 |
|
Newbie
Joined: 17 Nov 2005
Posts: 3
|
Came across a road block upgrading to WMQ v6 based what on the migration info specifying:
"Migrate the repository queue managers (that is, those queue managers holding a full repository) first, before migrating the other queue managers." under the clustering session
However due to scheduling conflict, the full repositories Queue managers will not be upgraded till later and that means it'll put a stop on all the other upgrades.
Anybody know how to get around this? |
|
| Back to top |
|
 |
| jefflowrey |
| Posted: Thu Nov 17, 2005 7:29 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
Some people had big problems when moving from 5.2 to 5.3, when they did not upgrade their repositories first. The addition of SSL parameters to channels in 5.3 corrupted the datastructures of the 5.2. repositories.
I'm not saying this will happen when moving from 5.3 to 6 - I don't know either way.
But you should reexamine the risks of the "scheduling conflict" and see if you can't upgrade the repositories first.
If nothing else, you may be able to change which QMs are full repositories without affecting your current infrastructure too much, and then upgrade your repositories first that way.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| schow |
| Posted: Fri Nov 18, 2005 6:56 am Post subject: |
|
|
Newbie
Joined: 17 Nov 2005
Posts: 3
|
Thanks for the suggestions.
We'll see what we can do with the scheduling. Too bad we don't know what will be the exact implications without upgrading the FRs first other than the gut feeling it's better to do that first so that we can make a more intelligent decision with the planning. |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Fri Nov 18, 2005 7:01 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
This is why a lot of shops have multiple environments, including a lab/sandbox type of setup. So you can find these things out in a safe way.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| schow |
| Posted: Mon Nov 28, 2005 7:20 am Post subject: |
|
|
Newbie
Joined: 17 Nov 2005
Posts: 3
|
| FYI in case anyone interested. Confirmed from IBM that there is no requirement to upgrade the full repos first unless we're working with IPv6 Qmgrs in the cluster or trying to exploit the new workload balancing enhancements, otherwise, we should be safe to proceed to upgrade the partial repos Qmgrs first. |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Mon Nov 28, 2005 7:30 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
| schow wrote: |
| FYI in case anyone interested. Confirmed from IBM that there is no requirement to upgrade the full repos first unless we're working with IPv6 Qmgrs in the cluster or trying to exploit the new workload balancing enhancements, otherwise, we should be safe to proceed to upgrade the partial repos Qmgrs first. |
In other words, the situation is the same as from v5.2 to v5.3 - as long as you don't use the new features, you won't corrupt the data stores of the repositories by putting new data into old data structures.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
|
|
