Re: Sending out trap when a process crashed

2011-08-12 Thread Dave Shield
On 10 August 2011 18:30, Phong Nguyen (ndphong) wrote: > My other question is what appropriate oids I should use for the trap. I > could see some pre-defined traps for link up/down, cold start, warm > start, but not for a process to be restarted. I'm not that I can think of a suitable pre-defined

RE: Sending out trap when a process crashed

2011-08-10 Thread Phong Nguyen (ndphong)
...@googlemail.com] On Behalf Of Dave Shield Sent: Wednesday, August 10, 2011 8:20 AM To: Phong Nguyen (ndphong) Cc: net-snmp-coders@lists.sourceforge.net Subject: Re: Sending out trap when a process crashed On 9 August 2011 18:09, Phong Nguyen (ndphong) wrote: > In my case, I'd spawn the killed proce

Re: Sending out trap when a process crashed

2011-08-10 Thread Dave Shield
On 9 August 2011 18:09, Phong Nguyen (ndphong) wrote: > In my case, I'd spawn the killed process again How are you restarting the killed process? Because that would be the natural point at which to generate the trap. If you've got a script that will restart the killed process, then simply add a

Re: Sending out trap when a process crashed

2011-08-10 Thread Wes Hardaker
> On Tue, 9 Aug 2011 10:09:54 -0700, "Phong Nguyen (ndphong)" > said: PN(> Could I use the DisMan Event MIB to generate the trap in my monitoring PN(> process? You can, but it'll be subject to similar delays. You'll need to tell it to monitor your agent with a particular frequency. If

RE: Sending out trap when a process crashed

2011-08-09 Thread Phong Nguyen (ndphong)
l.com [mailto:dave.shi...@googlemail.com] On Behalf Of Dave Shield Sent: Tuesday, August 09, 2011 12:31 AM To: Phong Nguyen (ndphong) Cc: net-snmp-coders@lists.sourceforge.net Subject: Re: Sending out trap when a process crashed On 9 August 2011 02:52, Phong Nguyen (ndphong) wrote: > I'd like

Re: Sending out trap when a process crashed

2011-08-09 Thread Dave Shield
On 9 August 2011 02:52, Phong Nguyen (ndphong) wrote: > I’d like to build and send out a trap when one of the processes that I > monitored got killed or crashed. The simplest would be to use the "proc" directive to monitor the process of interest, and the "monitor" directive (or "defaultMonitors

Sending out trap when a process crashed

2011-08-09 Thread Phong Nguyen (ndphong)
Hi, I'd like to build and send out a trap when one of the processes that I monitored got killed or crashed. However I could not find a suitable trap type and oid value for a meaningful variables_list. I could see the use of the HOST-RESOURCE-MIB process name/id/status, but also would like to have