Monitoring NIC errors

[bammy]

Hi everyone, I apologise if this is a fairly basic one.

I'm trying to use Insight Manager to monitor all our servers for any receive errors to the NICs. We're finding these do occur occasionally due to duplex mismatches.

From the reading I've done, I believe I may need additional MIBs, but I have absolutely no idea where to obtain them from?

Any suggestions or pointers in the right direction would be much appreciated.

Thanks

Chris

[Ferry Steenvoorde]

I guess your network requires all servers to be at least 100MB Full Duplex. Cisco switches perform much better when both the NIC's on the server and the ports on the switches are set at 100MB Full Duplex instead of Auto.

The information that you require is already available.

Open the HPSIM Console:
-Click Reports, New Report
-Select target servers or server group(s)
-Apply, Next
-Do NOT enter a Report Name (unless you want to save the query including the selections)
-Expand below 'Show Items to Show in Report': the 'General'-collection
-Expand 'Network Interface'
-Select 'System Name', IP-address', 'Speed', 'Duplex'.
-Select 'Layout' and 'Format' if required. The defaults will most times do.
-Click 'Run Report'.

I don't think you would require additional MIBs for this.

[bammy]

Thanks Ferry. That is very useful and it's shown us a couple to look at.

I know all the issues with Cisco kit and auto-sensing but our dept isn't responsible for the switches and we cannot guarantee things don't change. We also notice that some of the devices nailed down to 100/full are still receiving an awful lot of receive errors. This is the exact reason why I'm trying to get HPSIM to alert when it sees a server with increasing receive errors.

Any ideas?

[Ferry Steenvoorde]

You probably have to experiment a bit.

-Options, Events, Automatic Event Handling, New Task
-Create task 'with event and system attributes that I specify'
-Next
-Enter Task Name
-Next
-Select Events where 'Event Type' is 'Proliant NIC Events'
-Select the traps you wish to check. You would probably have to test what kind of 'Status Traps' are exactly checked. This part is not quite helpful.
-Next
-Select Systems (just what you like)
-Next
-Select Actions (send e-mail, write to system log?)
-Next
-Select time filter (leave default)
-Next
-Verify settings, and click Finish.

[bammy]

Thanks again but unfortunately I don't think receive errors by default generates any kind of trap from the client.

[bammy]

Any further suggestions anyone?