-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Brian Reichert

I have found that for a heartbeat v1 one-system node, you cannot have
auto_failback set to off (or on); heartbeat never really completely
starts.

I leave auto_failback out of the ha.cf file completely.  I think it will
log a warning, and default to 'legacy'.

See if that helps...

-- 
Brian Reichert                          <[EMAIL PROTECTED]>


Thank you very much Brian, that worked perfectly.  I emptied out my
haresources file and commented out the auto_failback from my ha.cf and
now I'm getting exactly the behavior that I was looking for.

But while I was working through this today, I also went a little ways
down the path of the "is Linux-HA even the right tool for what I want to
do?" question that I raised in my initial email.  I found the the
watchdog daemon [1] project on sourceforge and read up on it a little
bit (though I have not yet had time to compile it and try it out...
Working on it at the moment), and it seems that in addition to just
writing to the watchdog periodically, it will check a few conditions
such as free memory, temperature, etc. and can trigger the system to
reboot if any of those conditions reach a certain threshold.  Does
Linux-HA do similar checks with respect to the watchdog device, or does
it just write to the device, and if system load or screwyness
(screwiness?) happens to keep the heartbeat process from making that
write for a minute, the system reboots?  I see the value in the watchdog
integration in Linux-HA, but I wonder if the watchdog daemon might be a
better solution for my projects needs.

Thanks,
Michael

[1] - http://sourceforge.net/projects/watchdog/
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to