9.20 upgrade Issues: data_engine fails to start

[Daniel Blanco]

I'll start this thread now.  UIM 9.20 Upgrade Issues
=======================================

So in my lab decided to try out this new release and so far off to a great start... 
I reverted back to my 8.51 UIM image and then did the 9.02 upgrade. That went sort of well. Everything was running after a few stop/starts of services. I then upgraded to 9.20 and ok lets begin:

Probes and versions at 9.02 before upgrade to 9.20:

 

 

Started install at 9:26AM. Install finished 9:45 with many probes in INACTIVE state:

 

Then the usually nice "Welcome to 9.20" webpage failed to load…  :-(

 

 

Task Manager I see all new java exe's icons:

 I opened IM and tried to view the Archive but then IM froze:

Looked at the primary hub robot:

 

data_engine not running is always great…

Log:

Aug 21 09:33:49:212 [7400] 3 de: Before nimUnRegisterProbe

Aug 21 09:33:49:333 [7400] 3 de: Before nimSessionFreeList

Aug 21 09:33:49:334 [7400] 3 de: Before CleanUp

Aug 21 09:33:49:334 [7400] 0 de: ######################## STOP DATA ENGINE #########################

Aug 21 09:33:49:334 [7400] 2 de: nimEnd

Aug 21 09:33:50:003 [7384] Controller: Max. restarts reached for probe 'data_engine' (command = data_engine.exe)

 

I then did a full Nimsoft Robot stop, start and then re-checked probes. Waited a min or two and data_engine was still broken: 

 Same error in log:

Aug 21 09:59:34:556 [3592] 3 de: Before nimUnRegisterProbe

Aug 21 09:59:34:559 [3592] 3 de: Before nimSessionFreeList

Aug 21 09:59:34:559 [3592] 3 de: Before CleanUp

Aug 21 09:59:34:559 [3592] 0 de: ######################## STOP DATA ENGINE #########################

Aug 21 09:59:34:559 [3592] 2 de: nimEnd

Aug 21 09:59:35:560 [3076] Controller: Max. restarts reached for probe 'data_engine' (command = data_engine.exe)

Can anyone suggest what to do next here? This is MS SQL backend DB. This is actually my 2nd 9.20 upgrade attempt and the same thing happened last time. First time same issue. So I reverted and thought it was a fluke but it happened again on 2nd try. 

#cauimdata_engine


​​

------------------------------
Daniel Blanco
Enterprise Tools Architect
Alphaserve Technologies
------------------------------

[David Michel]

Dan, a team member reported this and really there has been no time for me to review so just passing on what might be a solution. 
use with caution and understanding. 

in UIM 9.2.0 there has been a change made to the QOS_DEF_ID TYPE for the QOS_POWER_STATE metric.
The type was changed from 1 to 0.
This was confirmed by running the following query: SELECT QOS_DEF_ID,TYPE,R_TABLE,H_TABLE,D_TABLE FROM S_QOS_DEFINITION where name='QOS_POWER_STATE';

then updated the TYPE from 0 to 1 using the following query: UPDATE S_QOS_DEFINITION SET TYPE = 1 WHERE QOS_DEF_ID = 881 After this change the data_engine queue started to process again.

------------------------------
Support Engineer
Broadcom
------------------------------

[Daniel Blanco]

Hi Dave,
Thanks for the reply. So I just ran the query part not the update:

Should the 'fix' query for me be ran on QOS_DEF_ID = 178  instead of your 881?

------------------------------
Daniel Blanco
Enterprise Tools Architect
Alphaserve Technologies
------------------------------

[David Michel]

yes

------------------------------
Support Engineer
Broadcom
------------------------------

[Daniel Blanco]

Thank you Dave. That fixed it...
Definitely need to put that on say a Known Issues section...    :-)


------------------------------
Daniel Blanco
Enterprise Tools Architect
Alphaserve Technologies
------------------------------

[Yu Ishitani]

Hello, Dan. Thank you for sharing, very much appreciated. Article (136125) posted.
https://ca-broadcom.wolkenservicedesk.com/external/article?articleId=136125