Windows server maxed out CPU and RAM, Notifications going crazy

Started by sharpspro, April 04, 2015, 02:35:23 AM

Previous topic - Next topic

sharpspro

I have NetXMS running on a Windows Server 2008r2 and having an issue since I updated to version 2.0-M1.

1. What seems to happen after a day of running, it will be maxed out on RAM and CPU. When this happens the notifications go crazy constantly poping up in task bar. Notification shows ("NODE NAME": Node status changed to WARNING) it will say warning, minor, major or critical and flashes quickly.

2. Within the NetXMS console the HEAP size at the bottom is jumping up and down.

Any ideas on this? Will updating to 2.0-M2 fix this issue? Looks like the change log shows that this version addresses a NetXMS crashing on some versions of Windows.

Thanks in advance!! :)


Igetyaall

Hey what kind of database are u using ?

Sometimes MySQL  and nextms dont get along to well ^^ , try to reset the Database ...

Alex Kirhenshtein

Notification are most likely due to the fact that processing inside NetXMS is more or less stalled (due to heavy resource usage).

Heavy resource usage, however, is highly unexpected. Can you please provide more details about your setup? Amount of memory, number of cores, database engine (local / remote).
Please also setup memory consumption monitoring for processes "netxmsd.exe" and "nxagentd.exe" (DCIs "Process.VMSize(netxmsd)" and "Process.VMSize(nxagentd)") and post screenshot of line chart for last 24 hours.

2.0-M3 will be publicly available in the next few days, you might want to try that.

sharpspro

Just want to share how I got my issue resolved. I created Tabled DCIs on a template that applied to all server nodes. I created Tabled DCIs for FileSystem.Volumes, System.InstalledProducts, System.Processes and System.Services.

So I removed the Table DCIs from all nodes it was applied to and have not had an issue since. I think I should have changed the polling time from 60 sec to something less frequently, since its pulling so much info each poll.

The CPU is back to normal as well. I liked the Table DCIs but not sure if this is ready or I just configured it wrong.

Thanks!

Alex Kirhenshtein


sharpspro