Categories
- All Categories
- 15 Oracle Analytics Sharing Center
- 15 Oracle Analytics Lounge
- 208 Oracle Analytics News
- 41 Oracle Analytics Videos
- 15.7K Oracle Analytics Forums
- 6.1K Oracle Analytics Idea Labs
- Oracle Analytics User Groups
- 76 Oracle Analytics Trainings
- 14 Oracle Analytics Data Visualizations Challenge
- Find Partners
- For Partners
How to debug/troubleshoot issue related to NQSGETUSERS in OBIEE 12c ?

We are facing issue related to OBIEE 12c login after looking into the log, we found
UserPopulation: GetUsers took 5 seconds for SQL {call NQSGetUsers
Now, we are not sure about the flow of this method execution, in other word where to look or debug ? What might be causing NQSGetUsers execution longer. We don't see any documentation about this method and what are related components.
So, anyone has any idea on what NQSGETUSERS does ? and how to debug this method (in other word, what are all components/layers I should verify)?
Answers
-
-
Yes, of course. I don't see this post answer anyway what I asked. My question is very specific to how NQSGetUsers works and what are related component/layer I should check ?
0 -
1.) Which "12c" version? We are standing at nineteen (19!) "12c" versions?
2.)
3064279 wrote:We don't see any documentation about this method and what are related components.
And you won't because it's one of the internal ODBC calls used inside the OBI Server.
3,) Looking at the other thread - this change here didn't affect you at all?
4.) Did you check the responsiveness of the RCU DB and tables?
5.) Did you do the testing with the init blocks which are also mentioned in the other thread?
0 -
1.) To be very specific, it is 12.2.1.3.0
3.) Not exactly
4.) You mean, directly connecting them through some SQL client and check their performance ?
5. ) So, I should check/test all the init blocks ? There are 25-30 in my case.
0 -
3.) Adam said that using SIDs instead of service names eliminated a lot of issues - personally I'd at least *look* at this
4.) Yes
5.) Yes
Performance issues aren't something where you can set GO_FASTER=true. They do actually require work and effort on your part.
0 -
Just to mention, we have more than 5 machines and all of them are using service_name instead SID, but we are having this issue in 2 machine. I checked your other suggestion, where you mentioned about bad RCU, I would like to know what do you mean by bad RCU ?
Also, right now we are debugging this issue by disabling and enabling init block to see, if init block has anything to do here.
0 -
You can also enable the tracking of initblocks in Usage tracking table s_nq_initblock.
See Welcome to OBIEE12c - Set-up Usage Tracking – Part 1 - Red Stack Tech for an example.
0