The secondary machine should have the same hosts.cfg and analysis.cfg. The primaries alerts.cfg should be copied periodically (whenever changed), to the secondary as something like alerts.cfg.primary.
Then the secondary has an alerts.cfg file that has one entry in it. And that entry just watches the primary:
HOST=xymon1 COLOR=red SERVICE=bbd,conn DURATION>2 SCRIPT /usr/local/scripts/xymSecondaryTakeOver $NG-IM-ONCALL MAIL $NG-IM-ADMIN
The script needs to copy the primaries into place.
Of course you’ll need another entry in the primary’s alerts.cfg file for the secondary to release when the primary is back online.
HOST=xymon1 COLOR=red SERVICE=bbd,conn SCRIPT /usr/local/scripts/xymSecondaryRelease $NG-IM-ONCALL MAIL $NG-IM-ADMIN
Which copies the secondary alerts.cfg file back in place.
That’s the basics. Other things to think about, is the two talking to each other when you disable or acknowledge alerts. Are your monitored systems going to send to both machines or just one, and the one works as a proxy. How you keep files in sync, etc.
In our system, the secondary is actually a proxy server. It can get to some systems the primary can’t get to directly. And there are a few systems the primary can get to that the proxy can’t.
Paul.
From: Xymon [mailto:xymon-bounces at xymon.com] On Behalf Of Eli Sent: Monday, May 05, 2014 4:35 PM To: xymon at xymon.com Subject: [Xymon] Xymon master - slave Setup
How I can setup 2 xymon server as slave max/master for fail over? Please advise. Thanks