We have a highly available environment for layer WebApp.
In total there are 4 servers.
A server with AIX OS layer for Essbase / Ordbms
A Windows OS server for layer Else
And two servers with Windows OS for layer WebApp
WebApp servers share the following applications:
Foundation (Shared Services and Workspace)
The problem occurs in the work in the Calculation Manager, any changes or creating new rules are not seen by all users. Only some see the changes
To verify the information in the relational database table specifically CALCMGRRULES and information exists.
At first we thought the problem was the load balancer. But now I do not believe that is the problem and we perform the following tests:
1. - Tells the Load Balanced is only allowed to work on a single server and always the same. For this hyperion services were operating on the two servers.
Result = the problems continued
2. - Hyperion services were down one of the servers.
Result = The problem is solved
There is someone who knows how to solve the problem?
I need to work with the two servers.
Calc Manager caches all the objects that is in the RDBMS when it starts up. So when you have 2 web app configured for calc manager, both have their own copy of these cahced objects. So when a user modifies an object that hits one web app, the changed information is propogated to the other web app through a SQL table CALCMGRCACHEEVENTS. So the server one write information of an object that is changed to this table (add,modify or delete) and the other web app(s) periodically reads from this table to refresh the cached copy of objects. By default the refresh rate is 10 seconds (so you won't see the change for 10+ seconds on the second server).
So from the problem description, it appears either the second web app is not reading from this table or the refresh is very infrequent that it does not pick up the changes quick enough from the other server.
By default this propogation is turned off. This is only turned on when the web app starts up and detects that there are multiple calc manager web apps configured.
Are you on the latest patch? (Patch 104 had been out for some time). If not, please apply the patch 104 (it is cumilative so need to only apply this patch) and see if the issue persists. Makes ure that browser cache is cleared after applying the patch. If this issue still exist, please file a SR and we will have to get on a call and check what is happening.
That is correct. If you do have Planning server running on a box that is different from Calc Manager, you should stop Planning and copy the jars from <ORACLEHOME>\common\calcmgr\188.8.131.52\lib on the calc manager server box to the planning one.
Planning on both servers I have and CLC. This would cause a problem?
What specific changes will be to apply the patch 104?
There is some risk to the platform?
Be reversed if presentarce a problem?
The instructions on how to apply the patch is in the zip file that you download. As per our internal testing and the response from the clients who have applied the patch, we feel it should be pretty safe to apply the patch.You should also be able to undo the patch if it poses an issue for you.