Chris Pepper wrote: > We have a couple small HPC compute clusters, and would like to monitor > our nodes. They aren't large enough to justify their own Nagios > installations on the head nodes, and the heads aren't particularly > trusted in our network topology. > > But we would like to monitor health of the compute nodes, on a Nagios > server which *cannot* connect to the nodes directly. > > I didn't find this in the wiki, SF, or Exchange, and it doesn't look > like something I could do with a single NSCA. > > Is anyone doing this, or forwarding/tunneling Nagios traffic for > another reason? I'm considering running a couple ssh tunnels on the head > node, pointing back to the monitoring servers, but not sure how well > this will work. > > Suggestions welcomed. >
I think I'd solve this using a small custom script that runs all the checks you want against the nodes (I suppose all nodes require more or less identical checks) and sends the results back to the Nagios server as passive checks. If the head nodes aren't allowed to talk to Nagios, they could publish the checkresults (along with a timestamp) through some other means, like http, ftp or even just a simple netcat session where a polling script on the Nagios server can fetch them later. Make sure to include a timestamp in the results-file if you do that, so you can verify that the checks are actually being run. Interesting problem. I'd take it kindly if you keep us posted :) -- Andreas Ericsson andreas.erics...@op5.se OP5 AB www.op5.se Tel: +46 8-230225 Fax: +46 8-230231 Considering the successes of the wars on alcohol, poverty, drugs and terror, I think we should give some serious thought to declaring war on peace. ------------------------------------------------------------------------------ _______________________________________________ Nagios-users mailing list Nagios-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/nagios-users ::: Please include Nagios version, plugin version (-v) and OS when reporting any issue. ::: Messages without supporting info will risk being sent to /dev/null