RMA on Windows NT 4.x SP6a failure

General chat about HostMonitor (all messages posted before March 07, 2003 available here).
Locked
User avatar
hmo
Posts: 81
Joined: Fri Feb 14, 2003 6:00 pm
Location: Denmark
Contact:

Post by hmo »

Hi Alex.

I have a couple of RMA's running on Windows 4.x SP6a servers (Oracle). RMA is installed as service. From time to time the service stops - no errors seen in eventlog - it just stops! When I try to start the service, it says that RMA "port" is busy (or used). Then I shut down RMA with a Process killer local on that server- and restart the service. Then everything is fine for 2-4-6 hours... and then we go again!

Another turn of this is, that I can use the RMA to monitor CPU, Performance Counter/s, Services, free diskspace etc. The only thing that fails, is the "Process" tests. It says: Unknown - 301 Cannot retrieve information.

I have updated the PDH.DLL on the server.. but no change happend!

Have You any clues about this?

Cheers,
Hans Mosegaard

PS: It is the same when I run the RMA as application (ie. not as service).

<font size=-1>[ This Message was edited by: hmo on 2003-08-04 06:37 ]</font>
KS-Soft
Posts: 13012
Joined: Wed Apr 03, 2002 6:00 pm
Location: USA
Contact:

Post by KS-Soft »

About problem with service: as I understand you have installed RMA on different systems but problem appears on NT 4.0 SP6 systems only? Do you use the same set of tests on all systems? How many tests and how often RMA performs the tests? I am interesting because I will try to reproduce this problem..

About problem with "Process" test: it does not use pdh.dll. CPU Usage test is working? Its strange because Process and CPU Usage tests need the same permissions. Do you use both tests to retrieve information from the same system? It is the same system where RMA is running or its remote system?

Regards
Alex
User avatar
hmo
Posts: 81
Joined: Fri Feb 14, 2003 6:00 pm
Location: Denmark
Contact:

Post by hmo »

Yes. I have installed RMA on a NT 4.0SP6 system. I perform about 55 tests every minute. I'we only seen this "service-stop" on this NT 4SP6 system (it's running Oracle).

CPU test, Performance (ie. Available Mbytes) is working... but check Processes does not. It says: 301 cannot retrieve information in the HM reply field! I check this through the RMA installed on that (same remote) system.. and YES :smile: it is the right one!

I have changed the RMA to run as an application (not service). No change to Processes tests... still not working!

When I go into the HM test, and try to choose dropdown list on Processes (ie. Alert when less than x <processname> is running".. nothing happens.

If I change back to direct RPC, ie check against the server directly from HM (like the good old days) :smile: - then reply is: Unknown. The process test I'm trying to retrieve is in fact active... I can see the name in the "task manager" on that server!

Do You have any "features" like a small investigation program to run locally on that system ... just to test if RMA in fact recieve- and have the right information/rights.. or some clerky local process test?

Cheers,
Hans Mosegaard

KS-Soft
Posts: 13012
Joined: Wed Apr 03, 2002 6:00 pm
Location: USA
Contact:

Post by KS-Soft »

>Yes. I have installed RMA on a NT 4.0SP6 system. I perform about 55 tests every minute. I'we only seen this "service-stop" on this NT 4SP6 system (it's running Oracle).

Ok, I setup RMA to perform 100 tests per minute (CPU Usage, Performance Counter, Process test methods) on NT 4.0 system. For now it works, will see what happans.
Unfortunatelly we don't have NT 4.0 with Oracle server but probably this problem does not have any relation to Oracle.

>CPU test, Performance (ie. Available Mbytes) is working... but check Processes does not

Did you play with Extensible Counter List application? Most likely you just disabled perfproc.dll

Regards
Alex
User avatar
hmo
Posts: 81
Joined: Fri Feb 14, 2003 6:00 pm
Location: Denmark
Contact:

Post by hmo »

Hi Alex.

Since I had trouple with the Process tests I disabled those tests yesterday. When those tests are disabled, the RMA service keep running okay.

Your most likely right about the relation to Oracle.

I did'nt play with Extensible Counter List, but I have now checked the local server setup with EXCTRLST.EXE - lots of strange things found. Not able to correct those when the Oracle Server is running, and a reboot did solve the problem for a few hours. I think it's a "local problem" for that Oracle Server.. so we cannon "blaim" HM for that one! You can stop the "test" at your end.

Cheers,
Hans Mosegaard
KS-Soft
Posts: 13012
Joined: Wed Apr 03, 2002 6:00 pm
Location: USA
Contact:

Post by KS-Soft »

Yeah, may be some problem with Oracle. RMA still running fine on our NT system, already performed 83,000 tests..
Will keep it running for a while

Regards
Alex
Locked