OBIEE Presentation Service unable to start by a batch job

Question

We are using OBIEE 10.3.4.1 in clustered environment with two BI servers and two presentation servers. We bounce the all the services on daily basis using windows batch job. Presentation Service on one of the bi presentation servers is failing with the following information in the Windows Event Log (We don't see any specific log in presentation server logs). When we start manually then it starts without any problems. It is happening only in production that's why it is critical. Please advise ASAP.
Event Log:
Event Type:     Error
Event Source:     ¨ô     le BI Utility
Event Category:     None
Event ID:     21
Date:          2/17/2011
Time:          5:20:43 AM
User:          N/A
Computer:     Z02FDCVSC28
Description:
The description for Event ID ( 21 ) in Source ( ¨ô     le BI Utility ) cannot be found. The local computer may not
have the necessary registry information or message DLL files to display messages from a remote computer.
You may be able to use the /AUXSOURCE= flag to retrieve this description; see Help and Support for details.
The following information is part of the event:
Attempted to use an unconstructed lock.
Current Owner thread id: 0 (0x0)
Current thread id: 5672 (0x1628)
Lock pointer: 0x6F0B291C
Lock created in file: d:\views\nightly\Tue\101341\Windows\vobs\090414.1900\analytics\server\include\Utility/Generic/SUGRecursiveLock.h
Lock created on line: 34
Lock created in function: NQRecursiveLock<class NQCriticalSection>::NQRecursiveLock
Error in file: .\NQThreads\SUGCriticalSection.cpp
Error on line: 242
Error in function: NQCriticalSection::Acquire
Thank you,

Answer 1

IMHO you shouldn't be restarting services on a daily basis. The only reason I can think off you will be doing that is to hide away real issues, such as memory leaks or stability issues. We run OBIEE in Linux and we have no need to restart services at all. In some cases the services have been running for several months without failing. We do use the autorestart option which means that the process will restart automatically should it crash, which sometimes happens. Unfortunately this option is not available on Windows, perhaps it is testament on the robustness and the Production readiness of that OS?
I would think your issue is either due to the account you are using to run the scheduled batch job or a timing issue where you need to give the service a few seconds between stop/start to clear all file handles. But still it's very bad practice to do what you do, especially if you are running in cluster mode which should give you HA.