[CentOS] Keepalived - spurious failovers

Wed Nov 12 16:33:59 UTC 2014
John Horne <john.horne at plymouth.ac.uk>

On Wed, 2014-11-12 at 11:12 -0500, m.roth at 5-cent.us wrote:
> John Horne wrote:

> >>
> > They are both virtual servers - so no UPS. Failover communication is
> > over the network.
> 
> Um, bingo: are the host systems on UPS's? What happens on the *host*
> systems at 03:56? They don't, perhaps, take snapshots of the guests then?
> 
No, no snapshots are taken. As said this is a spurious event which has
happened at 03:56 for the past two nights. However, we ran for a few
days before then with no problems. Before that it happened at something
like 5AM. It does not happen every night, nor at the same time
(usually).

I have set up a couple of cronjobs to check ifconfig and ping the
interface every few seconds. I also have a job that will monitor the
main interface for VRRP traffic since that should show what the priority
value is when a server claims to have received a higher priority from
another server.




John.

-- 
John Horne                   Tel: +44 (0)1752 587287
Plymouth University, UK