how to decrease frequency of SNMP trap polling by CIM?

[Geerts Nick]

We have 1 CIM7 management server and 1 server with the SQL database running. We noticed that CIM generates following select query every second:

"SELECT trapItems.trap_id, trap_name, trap_enterprise, trap_generic_id, trap_specific_id, trap_severity_id, flags FROM trapItems,trapItemsEx WHERE trapItems.trap_id = trapItemsEx.trap_id AND trap_generic_id = 4 AND trap_specific_id = 0"

We think SNMP trap polling from CIM generates this query. This puts however a serious load on our SQL server that also hosts other SQL databases of our enterprise. Is there a way to fine tune this polling? We tried to decrease the polling via the GUI on the CIM management server (Settings/Protocols/Snmp settings) but this does not change anything. Does anyone has a solution for this issue?

[Jadrice Toussaint]

Hello -

Go to Devices -> Tasks.

Edit the "SNMP Status Polling for Servers" task by clicking on the little pencil next to it.
Click on Next. On the next screen, there is a section that reads "Selected Query and Schedule". That is where you can change the frequency as to how often this task runs. The default is set to 5 minutes. Change it to whatever you like. I've attached a screenshot for you to see.

[Geerts Nick]

Jadrice,

I already tried that but this does not decrease the frequency of the select query (we ran an SQL profiler trace to monitor the SQL server). I also tried to fine tune other running polling tasks but without results. only when you disable the SNMP protocol from the settings page in CIM, the select query disappears but of course this is not a solution to our problem since we want to continue using SNMP. Do you have other suggestions?

[Jadrice Toussaint]

Did you decrease it just for the SNMP status for servers polling task or did you also decrease it fo rth other polling tasks?

[Geerts Nick]

I did it for all polling tasks but the frequency of the select query does not seem to decrease.

[Scott Shaffer]

The answer is no, there isn't anything you can do to limit this. The system does this automatically so that it can properly update itself and send out pages or emails.

This is the first time anybody has complained that this was causing trouble for them though. What issues are you seeing? This is a pretty light-weight query that should execute very quickly. Is it just that it happens once per second?

It is also possible that Insight Manager 7 isn't something that should be added to the particular database server in question - that the performance characteristics of IM7 conflict with other things you are using it for. This could mean that you should consider using a different or seperate database server for Insight Manager 7.

[Patrick Burns_1]

I would agree with the last post, you may need to move the dB to another box. In addition, MS SQL provides several ways to check the intregraty of a dB. You may want to run some DBCC commands against the IM dB. First one I would recommend is the DBCC CHECKDB. This of course should be run off-hours and you probably want to shut down the IM services first.