Re: heartbeat and monit v4

[Jan-Henrik Haukeland]

Marco Bisioli <address@hidden> writes:

> hi!
> I'm experimenting with monit and heartbeat as described in the howto
> (with nice failback), I have the following problem (for the moment I
> solved it using a ramdisk):
> 1) the active node fails, the second takes over
> 2) after a reboot, the failed node comes back, monit looks at the state
> file and starts all the services (even manual ones) as they were running
> before failure. This way services are running on both nodes (=problem).
> I simply put the state file on ramdisk so that state is "reset" between
> reboots.
> Is there a better way to do this?

Putting the state file on a ramdisk is not a bad idea, another idea
could be to remove the monit.state file in a rc-script[1] called at
boot time and before monit is started. There is no easy way to solve
this problem from within monit.

Thanks for the problem description, this problem and soultion should
be added to the doc.

[1] Not the rc script used to start monit.
-- 
Jan-Henrik Haukeland