tags 338391 = confirmed unreproducible thanks hi steve,
On Wed, Nov 09, 2005 at 07:17:58PM -0600, Steve Greenland wrote:
> What happens: I often (every few days? weeks?) find multiple nagios
> processes running -- not children, but independent master daemons.
> While this occurs, I get multiple notifications for events (or at least
> some -- I told you this would be useless). Running '/etc/init.d/nagios
> stop' will kill one, while the other must be killed by hand, presumably
> because the pidfile is wrong/missing.
i've noticed this problem as well in some cases, though i haven't been
able to root out the cause or reliably reproduce it myself. i think
this is the first time i've ever had to tag something both "confirmed"
and "unreproducible" at the same time.
my hunch is that the problem is most likely being caused by some
fork/exec failure in the nagios daemon not being properly
handled while executing some plugin. though this is only a hunch.
i'd be curious to see if we could find some more info
on the problem.
for example: when there are duplicate processes, is the older or the
newer in the pid file? which one(s) has an open filehandle on
the nagios.cmd socket? the next time this happens for either of us,
let's record that information.
sena
--
signature.asc
Description: Digital signature
