We are using BOXIR3 and are having 4 nodes in cluster. I want to remove one of the nodes. Can you please suggest me a way to remove the node. Is using serverconfig.sh the correct approach?

I use a main method to invoke my logic.

I get a session manager object and then enterprise seesion. It works fine for the first time but second time it throws an exception after getting the session object only. My theory is that since crystalenterprise returns a singelton, it is not able to return a new session object the second time.
Is there a way i can get a new session object every time i connect to a cluster like same as in .net sdk where we do sessionManager = new SessionManager.

I have 2 clusters with 3 servers each. I am using java sdk to logon to one cluster at a time. I logon to first one fine and then log off but then it dosent allow me to logon to 2nd one and gives an error that CMS service not found. If I do the other way around, it again lets me logon to 1st server but fails for the second one. I did some research and found that the issue is with the session oblject because it is a singelton and does not work with the second cluster. Can you please suggest a way I can logon to both servers one by one using java sdk.

So, my understanding is that you are upgrading in place your database and your FRS (input/output). In this case you really don’t have many options and clustering will happen automatically during the installation as long as you provide the correct inputs.

I have not done an upgrade of an XI R2 database to XI 3.1, but this is what I recommend trying:

1) Get two test servers, 1 test database, and 1 test location to put your FRS. Make sure they are configured and running XI R2 successfully. If you can only get 1 test server this is OK.

2) Shutdown completely all BO services.

3) Wipe the test servers clean of BO.

4) Initiate the install of the highest full install of XI 3.1 available, I think Service Pack 4 is a full install (if not SP3 is). Point the install at your test database. Provide the proper password for Administrator. Complete the install.

5) Stop all BO services on this newly installed server. Using CCM rename the cluster from the server name to the name you desire.

6) Start all BO services on this newly installed server. Test, look around

7) You may need to run “update objects” in order to obtain the XI 3.1 versions of the auditing reports and other BO-provided reports and universes.

8 ) If all is well, then leave this running, but shutdown all but the CMS.

9) Go to the second test server, if there is one, and install the same full install as you did to server #1. Provide the database info and Administrator password as before. You should see that the installer takes care of setting up the cluster because you are pointing it to a database where a XI 3.1 CMS exists.

10) After both servers are on the same patch version then you can advance 1 first to the next desired FixPack level and follow with the other.

Please share your thoughts on this. Of course, I would review the BO documents on this. In place XI 3.1 upgrades are not my forte as I have always had the luxury so far of moving to new servers and databases.

BTW, since you actually have two servers, another alternative is to do a migration upgrade. If you can get a new production database then you could basically remove one server from the current system. Install BO XI 3.1 to it and the new database. Then execute a migration from the old to the new using Import Wizard. Then when all is migrated, you could shutdown the 2nd server, wipe it and install XI 3.1 on it.

We have the BO XI R2 with two CMS server with clustering concepts.Everday more than 400 jobs will run from this server. sometimes some of the daily reports are not ruuning.if i stop any one CMS server then its running. Just i would like to know the reason for these kind of issue?