Global Ping Timeout Value

[Oran Turner]

We're running IM 7 SP2 with around 100 mixed Compaq/HP ProLiant servers. It's been a very reliable system, consistently keeping 3 administrators notified via SMTP to pager/cell. However, our single biggest complaint is what we consider unnecessary critical device notifications due to missed pings, as a result of a server reboot.

I'd prefer to set a global ping timeout of 3 minutes for all systems, before a critical device notification is sent. Is this possible? I'm aware of the Retries & Timeouts section under Automatic Discovery. Is this what I need to change? Keep in mind that we are not using automatic discovery, only manual discovery.

Thanks!

[Jadrice Toussaint]

Oran -

Modifying the ping timeout is not possible on the current version of IM. This was implemented on the new version that was released recently.
There is a way around it however, if you don't want to be notified for device status change, you can modify the snmp status polling for sever then uncheck the "Send status change events" checkbox or disable alerting for that device right before a reboot.

[Darrin Rawls]

You are right Jadrice. What you might want to do is just change the SNMP Status Polling task to run every 4 minutes or so. What happens is that before the SNMP polling task is kicked off, it does a ping to make sure the device is there; no reply and the device automatically goes critical. changing the time period for the polling task should solve your problem.

Remember, the polling task is just around b/c SNMP trap delivery is not guaranteed. That's a "fail-safe" mechanism to notify you of a degraded system if a trap was blocked somehow or the nic was down or something like that. Changing the polling time to 5 minutes or more should be okay for most systems; heck some you might only want to poll once a day or something (non-critical file/print for example). You can have multiple polling tasks tied to different queries....