topic Re: Hardware Status polling.... in Server Management - Systems Insight Manager

Hardware Status polling....

John Kramer — Fri, 26 Dec 2003 13:53:55 GMT

I have setup a task to send all reachable/unreachable events to my email and pager, however I would like to setup the timing so that normal 4-6 minute reboots do not get emailed but outages greater then say 10 minutes get emailed and paged. I have set the time out values on the Hardware Status polling task to 120 seconds with 3 retries. but this does not appear to have any effect. As soon as the Hardware Status poll runs it detects that the system is down and reports it. Do these timeout values actually do anything? If so can someone provide a details explaination, whats in the user guide is very limited.

Re: Hardware Status polling....

James Kennedy_4 — Mon, 29 Dec 2003 07:41:04 GMT

I have the same problem. I'm looking for a treshhold setting of some sort. We do reboots a lot and I really don't want those to be paging everyone that the server is down. Let me know if you find it, I'll keep hunting.

Re: Hardware Status polling....

John Kramer — Wed, 31 Dec 2003 12:34:25 GMT

Anyone from HP here this week, I would think the developers would be all over this forum with the release of the new product? Anyway this is what I would like to do if possible....I have 3 classes of servers, 1 mission critical, I need to know whenever one of these is rebooted. The default settings work fine for this. Then I have a second class of servers that if one reboot occurs and the system comes back up in a resonable amount of time then I don't want it to send out at page. Then I have a third class (development systems) that I can be down for even longer but still need hardware monitoring.

Re: Hardware Status polling....

Mike Kapos_2 — Wed, 31 Dec 2003 17:23:47 GMT

I have the exact same problem.. Would be nice if someone could give us some ideas..

Re: Hardware Status polling....

David Claypool — Fri, 02 Jan 2004 12:03:59 GMT

Whether by ping (new in hpSIM) or SNMP poll, when a system is unreachable because the systems is truly down, is in the middle of a reboot or some part of the communications infrastructure between hpSIM and the target machine is down, hpSIM has no way of knowing. Lengthening poll cycles or setting long timeout periods does not guarantee that the task will not fire in the middle of a reboot and make the device go critical.

Re: Hardware Status polling....

James Kennedy_4 — Fri, 02 Jan 2004 12:10:38 GMT

David,
Is there any way to setup a threshold at all? So that after 1 unsuccesfull polling loop it does not consider it down?

This setting was in Insight Manager 5.5. It was called "Inaccessible Alarm Retries". Setting this to 3, it would go through the loop 3 times before it actually considered the server to be Inaccessible. If SIM can't do this, then I guess we'll still be stuck using IM 5.5 for our monitoring.

Re: Hardware Status polling....

John Kramer — Tue, 06 Jan 2004 11:18:57 GMT

Does SIM have anything that would work the same way?

Re: Hardware Status polling....

Jadrice Toussaint — Tue, 06 Jan 2004 16:18:17 GMT

hmm I tought his feature was going to be added to SIM but it was not. My recommendation to you is to run WhatsUp Gold for along with SIM. WhatsUP Gold allows you you to set polling triggers and you can have it set to alert you after the device misses the number of polls specified. Unfortunately SIM lacks in that department.

Re: Hardware Status polling....

Darrin Rawls — Tue, 06 Jan 2004 17:12:44 GMT

There are actual retries already built into the code, but ah yes, we did not put a GUI around tweaking this.

If you change the Status Polling task to run every 7 minutes (if you want an interval of 1-6 minutes for a reboot), then you won't get the "false" notifications you are talking about. You will get an event when the server is down more than 7 minutes.

Re: Hardware Status polling....

John Kramer — Tue, 06 Jan 2004 17:20:04 GMT

If I change the Hardware Status polling task to run every 7 mintues and a server reboots after 4 minutes and is not back up it will generate a status change correct?

You said there is no GUI but, is there a CLI interface to change the polling retries?