IBM Connections – trouble adding additional nodes
we are currently involved in a project where we installed a 1 node IBM Connections 6 Cluster and later added a second node to the cell.
So far so good… Everything that needs to be done after adding the second node is described here … Everything? Yes, mainly… but not in the mandatory details as I think! Missing custom properties and other settings might result in non-functional nodes. Especially not setting httpSessionIdReuse results in strange errors – IBM makes it themselves easy by saying:
I would expect some more details, as you do not really know what exactly custom properties have been set…
There are also other steps you need to do to set all properties – I tried to summarize everything:
- Servlet and command caching – this is activated on Node1 and shall be activated on Node2 too (It is recommended to disabled this for the Files application when you use the large deployment and plan to use large file downloads)
2) Log settings need to be adjusted to match node 01
3) WebContainer custom properties need to be set
4) Amount of WebContainer threads needs to be adjusted
5) JVM custom properties need to be set
6) HttpSessionIdReuse has to be set for all AppServers + additionally session timeouts …
If you have a large deployment where we talk about 18 servers – it is a hell of work to do all this manually. We created a small script – I will try to post this as soon as we had time to make this a bit “nicer” …
In an earlier post of Nico Meisenzahl he described that after adding a node to an IBM Connections 5.5 Cluster search did not work. It seems that this has been solved in IBM Connections 6. Even though IBM tells you to create WorkerManagers for the second node, I think this is not mandatory as in IC6 all WorkerManagers are created within the Cluster Scope… So thoses workers are valid for all AppServers in the cluster. The search started working directly after adding the node without creating additional Worker managers.