My client is looking for a design which includes providing failover for the deployment manager. How do other companies handle this? I hear little talk about this subject and would like to know if anyone has had experience with such a set up?
I understand the deployment manager is not critical part of the WebSphere application doman as a whole and can even be turned down if not being utilized. And there are ways of administering to the application servers using wsadmin and the WAS shell scripts to, but the client is looking at options for providing failover.
We've proposed a solution whereby the config information for the deployment manager is stored in a Shared Storage location. The config is shared by 2 installations of deployment managers. One deployment manage is running at any given time. If one manager goes down, the other is brought up manually to take over.
The Primary DMGR is assigned a dedicated IP. Failover involves stopping the primary DMGR and unmounting the deployment manage profile partition. Then activate the IP on the secondary deployment manager and mount the profile partition. Then start the secondary deploymnet manager up.
In theory this should work, but I need to know what the standard practice is. Or whether anyone has attemped what I'm proposing and can give advice or opinions.