SM_server error in LMS2.6

Unanswered Question
Jul 24th, 2009

Hi,

I am using LMS 2.6

SM_server process is taking high cpu utilization. I am not able to configure, polling parameter or Managing Threshold as I am getting the error " PTMServer" process is down.

But the process status is "Running Normally" . I have also restarted the "PTMServer" process but no result.

Attached the DFM.log.

Attachment: 
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
Joe Clarke Fri, 07/24/2009 - 09:23

It appears sm_server is crashing due to bad trap data. Post a screenshot of DFM > Configuration > Other Configurations > SNMP Trap Forwarding.

Satya Siba Sund... Sat, 07/25/2009 - 03:51

Hi,

I havce stoped syslog service and DfmServer process after that also DBsrv9.exe process had shown high CPU utilization. I have above 3000 device monitored in my cisco works. Out of these devices 586 devices showing under confilcting device in DFM.

is the problem because of these confilcting devices?

I have attached the CPU utilization screen shoot of CPu utilization screen shoot when both syslog and DFM are running (cpuutil.jpg), snmp trap forwarding(trapforwarding.JPG) and DFM device summury (dupdev.JPG).

Please help me if this error is not because of DFM device conflict.

Thanks.

Attachment: 
Joe Clarke Sat, 07/25/2009 - 11:38

The problem is being caused by too much or bad trap data. However, you are managing too many devices in DFM. DFM 2.x could only handle up to 1500 devices in ideal conditions, and you are trying to manage twice that.

First, to prevent the crashing, temporarily disable trap reception under DFM > Configuration > Other Configurations > SNMP Trap Receiving. Then, reduce the number of managed devices to a supported value. While 1500 is a rule of thumb, the real number depends on the number of managed ports and interfaces. The real rule is:

DFM supports configurations of up to 45,000 ports/interfaces, of which 6,750 (or 15%) are managed. This support was tested with an average of 30 ports/interfaces per device.

You can use the NMSROOT/objects/smarts/bin/sm_tpmgr --sizes command to see the current number of managed ports and interfaces.

Once you have the managed device count under control, you should stop sending traps from devices no longer managed by DFM. Once you do that, you can try re-enabling trap reception in DFM, and see if the crashing goes away.

Actions

This Discussion