Re: www-commits list not getting commit emails

[Bob Proulx]

Andrew Engelbrecht wrote:
> I should note that this is the vm that has had an "UNKNOWN" alert for its
> mail queue length for 602 days, because there is a bug somewhere in check_mk
> on that vm or our Nagios server.

Hmm...  Good to know.  It would be good if a Nagios check for MTA life
worked.  I will put that on my list to nag about looking at sometime.
I don't see anything obviously different on vcs1 versus vcs0 for
example.  Both have check-mk-agent installed and both have firewall
rules allowing the connection to it.

I do actually look at the emailed noifications.  But I admit I only
rarely look at the web dashboard.  I should look at the web dashboard
more.  And the email notifications were that *all of everything* was
failing so I rather deleted all of them.  I should have looked at the
Nagios web dashboard.

For these systems a mail queue length is not as useful as an MTA life
check.  As there are often valid times to have a deep queue.  But the
MTA should always be running.

The storage array failure is a pretty unusual situation though.  It's
unlikely to be a repeating problem.  Although it would be possible to
monitor and automatically restart all of the daemons that normally is
not needed.

I thought about rebooting all of the nodes as a preventative but since
everything seemed to have been working okay I didn't do it.  However
if I had rebooted all of the nodes then that would have prevented this
problem.  As a just in case I will queue up a reboot of the other
nodes next week during daylight hours as a just in case prevention.

Bob

signature.asc
Description: PGP signature