We are running HM 4.42 on W2K SP4 with hundreds of alerts (we are not using RMA) checking at regular intervals. Within the last couple of months we have started seeing an issue where the HM process will consume all CPU resources on the server and in order to resolve the issue the HM task must be killed. When HM is restarted it runs fine for a while and then the CPU utilization issue returns.
Let me know if more information is required to begin trouble shooting this issue.
High CPU Utilization for HM 4.42
What about other resources? Handles, USER and GDI objects, memory?
I have tried to determine exactly what HM is doing that is using the entire CPU but have been unsuccessful. Do you have any suggestions for collection the data you are asking for?
If you stop monitoring (without closing application), CPU usage drops to 0% or not? If you resume monitoring, CPU Usage raises up to 100% again?
When the issue occurs we usually kill the HM process. During the next occurrence I will stop and restart monitoring to see if that helps.
What kind of test do you use? Ping? SNMP? URL...?
We use ping, port, url, nt log, service, freespace, cpu, IMAP and probably a couple more that I cannot think of right now.
I suspect that one of the IMAP tests that are monitoring mailbox space is causing the issue but I cannot disable it until I have more than a hunch that it is the culprit since the account it monitors is business critical.
I have tried to determine exactly what HM is doing that is using the entire CPU but have been unsuccessful. Do you have any suggestions for collection the data you are asking for?
If you stop monitoring (without closing application), CPU usage drops to 0% or not? If you resume monitoring, CPU Usage raises up to 100% again?
When the issue occurs we usually kill the HM process. During the next occurrence I will stop and restart monitoring to see if that helps.
What kind of test do you use? Ping? SNMP? URL...?
We use ping, port, url, nt log, service, freespace, cpu, IMAP and probably a couple more that I cannot think of right now.
I suspect that one of the IMAP tests that are monitoring mailbox space is causing the issue but I cannot disable it until I have more than a hunch that it is the culprit since the account it monitors is business critical.
You may use standard Windows Task Manager, it displays such information on "Processes" tab. Use menu "View"->"Select columns" to select additional counters (such as "GDI objects", "User Objects")Do you have any suggestions for collection the data you are asking for?
These test methods don't use 3rd party DLLs, we have checked them for resource leakage... Do you use "Performance Counter" test? Do you use ODBC logging?We use ping, port, url, nt log, service, freespace, cpu, IMAP and probably a couple more that I cannot think of right now.
Regards
Alex
wes:
Is is possible you've got some kind of report that is growing in an uncontrolled fashion? Or some kind of huge log?
Just to compare, we've got approx 3,000 defined tests, 100+ RMAs -- HM estimates load at 17 tests/sec. Running on a Dell 1650 (dual CPU) under Win2K, SP4+ -- most times CPUs are 5-10 percent utilized. Every few seconds a report or two will run that brings cpu up tp 50-60 percent but it drops down fairly quickly. (We are now running HM 5.02 beta but the above info was true last week as well when we were running 4.86).
On just one occassion a couple of weeks ago, HM began to display "Out of memory" dialogs. It had been running 3-4 weeks non-stop so we just rebooted and the issue has not occurred since.
Is is possible you've got some kind of report that is growing in an uncontrolled fashion? Or some kind of huge log?
Just to compare, we've got approx 3,000 defined tests, 100+ RMAs -- HM estimates load at 17 tests/sec. Running on a Dell 1650 (dual CPU) under Win2K, SP4+ -- most times CPUs are 5-10 percent utilized. Every few seconds a report or two will run that brings cpu up tp 50-60 percent but it drops down fairly quickly. (We are now running HM 5.02 beta but the above info was true last week as well when we were running 4.86).
On just one occassion a couple of weeks ago, HM began to display "Out of memory" dialogs. It had been running 3-4 weeks non-stop so we just rebooted and the issue has not occurred since.
The log.htm file was around 30MB so I set it to create a new file daily to see if that might be part of the problem.
We do not do any reporting from HM just HTL logging to allow for a historical record of events when necessary.
The system running HM is a Proliant 3000 (P3 550MHZ CPU & 256 MB of memory) running W2K Server SP4.
We do not do any reporting from HM just HTL logging to allow for a historical record of events when necessary.
The system running HM is a Proliant 3000 (P3 550MHZ CPU & 256 MB of memory) running W2K Server SP4.
Resource usage is resonable, don't know why it uses CPU so much 
Could you send your settings for testing to support@ks-soft.net? We need *.LST files, hostmon.ini and .HML file with tests.
BTW Do you have installed some antivirus monitor (particularly Norton Antivirus or McAfee) or personal firewall? Could you try to disable it?
Regards
Alex
Could you send your settings for testing to support@ks-soft.net? We need *.LST files, hostmon.ini and .HML file with tests.
BTW Do you have installed some antivirus monitor (particularly Norton Antivirus or McAfee) or personal firewall? Could you try to disable it?
Regards
Alex