RMA 301 Error - Startup timing issue

[timn]

This is a minor issue for us but thought I'd ask for your recommendations.

We have a bank of 60 servers that we are monitoring. One of the items that we are monitoring is Performance Counter \Web Service(_Total)\Current Connections. There is an RMA on each of the 60 servers.

Currently, all tests for a specific server are dependent upon a master (simple ping) test for that server.

Sometimes, we reboot all 60 servers in the early AM. After rebooting, some of the Perf Counter tests mentioned above return a result of "RMA: 301 - Error: The specified object is not found on the system." Usually, no more than 4-6 of the 60 servers have this issue.

They continue to report this error until the RMA for the machine is reloaded using the RMA Manager. After reloading, these test succeed and the value of the Perf Counter is returned.

So I'm thinking that this some kind of timing issue - i.e. HM is requesting the value of the Perf Counter before (in this case WWW Publishing Service) has fully loaded. Once the error has occurred, for some reason the RMA continues to report the error -- even though the WWW service is fully loaded and counters are available -- until the RMA itself is reloaded.

One thought I had was to create a new test on each of the 60 servers that checks to see if the WWW Service is responding, then make the Perf Counter test dependent upon that test -- i.e. so we only ask for the Perf Counter once we know the service is loaded.

Do you see a better solution?

[KS-Soft]

So I'm thinking that this some kind of timing issue - i.e. HM is requesting the value of the Perf Counter before (in this case WWW Publishing Service) has fully loaded. Once the error has occurred, for some reason the RMA continues to report the error -- even though the WWW service is fully loaded and counters are available -- until the RMA itself is reloaded.

Yes, its possible. When pdh.dll reports error, often it continues to report this error until reload.
Unfortunatelly application (RMA in our case) cannot reload DLL immediately because Windows requires 10 min timeout.

I think "External" mode should help to avoid this problem. RMA can use perfobj.exe utility to perform Performance Counter test. You may find perfobj.exe in HostMonitor's directory. Copy utility into RMA's directory, add "PerfWorkMode=3" line into [Misc] section of the rma.ini file and restart agent.
http://www.ks-soft.net/hostmon.eng/mfra ... m#MiscPerf

Regards
Alex

[timn]

Thanks Alex. I'll give this a try and let you know how it works.