4 people reporting the same thing on the same day...I hope Oracle Support is listening and have arw connecting the dots there
In our prod env as well we got same failure yesterday ! Rep variables stopped refreshing ....
From last April onwards we faced this issue almost 5-6 times ..
Over the time I found the easiest method to refresh them again. Just to share it is as below .. No need to update or restart ...
1. Open the rpd in online mode
2. Disable all dynamic init blocks in one go and save the rpd.
3. Enable all disabled init blocks in previous steps and save the rpd.
It's done ! ... all rep dynamic variables will be uptodate and auto refresh will become functional...
Agreed that works but this is such a central and critical functionality that it HAS to be solved by Oracle once and for all in the core product and not be a manual hack where someone has to go a actively do things by hand in a productive environment.
Is there an SR for this? Are Oracle even aware of this issue?
From my side multiple. Every time it happens a new SR is opened for this.
Also pointed to this topic in the SR that multiple sites have this issue.
If only there were some thing like sentiment analysis or cluster analysis or .... gasp .... a relationship graph which could point out multiple related occurrences of this issue!
But that's just crazy. Technology isn't that far advanced yet.
From my side we have also reported this issue multiple times to Oracle support and we have send them several diagnostic_bundles each time the error occured.
Regarding the steps of Onkar Nath Tiwari we almost do it the same way:
1. Open .rpd-file in online mode
2. "Test" the data source of the respective inilization block and commit
3. Save the .rpd-file
Yesterday I got an advice of an external consultant:
1. Create a new connection pool within Administration Tool
2. Set cache size to NULL / deactivate it
3. Set your inilization block to this new connection pool and only use it for your variables
All these hacks may work but are quite dangerous because you risk bringing/inventing in errors which weren't there before. It's just a manual crutch.
Re new connection pool that's actually the most scary one I've seen so far.
The key question is simply: how long until this first happens on a critical OAC instance and the pressure gets high enough that there is a fix in the codebase
We have raised an SR for this too, a couple of months ago
Just seen this from Oracle: Oracle Support Document 2338889.1 (OBIEE 12.2.1.X Reports and Dashboards Using Dynamic Repository Variables Not Refreshing and Showing Incorrect Data ) can be found at: https://support.oracle.com/epmos/faces/DocumentDisplay?id=2338889.1
That looks hot off the presses!
Is there a way to subscribe to new documents relating to a product that you are using, or were you just searching 'My Oracle Support'?
I've sent you a DM.
I'll be installing this patch in our Development environment today and will keep this thread informed of the progress. Thank you!
Thanks for spotting Joel!
I do find it strange to have this included in the solution:
SET VARIABLE DISABLE_CACHE_SEED = 1,DISABLE_CACHE_HIT=1
Also have asked Oracle support if this patch is going to solve this issue [as they did not reply to my SR at all]