[Maintenance] monit on rshg054
Luke T. Shumaker
lukeshu at sbcglobal.net
Fri Oct 4 17:50:47 GMT 2013
Earlier today I noticed that labs.parabola.nu was throwing a bad
gateway error (what's that, 503?).
So, I hopped on the box to figure out what was going on. First of
all, everything but transmissiond had been unmonitored by monit.
I first-ish did a few general troubleshooting steps, I added myself to
the monit group, and upgraded monit. The newer monit doesn't ship
with an rc script, so I manually re-installed the old one.
In most cases, it was stupid mistakes in the monit
configuration--looking for a PID or socket file in the wrong place.
In the case of parabolaweb, it was checking to make sure the socket
was talking HTTP--in reality, the socket talks FastCGI.
After a while, I got an email saying that it timed out Nginx
(restarted the service on 3/3 checks). I disabled the HTTPS check, it
seems to have been problematic, but Nginx really did lock up. I had
to kill -9 it, and remove the lock file. It seems to be fine now, but
I will keep a eye on it.
~ Luke Shumaker
PS: I have a few questions about monit CLI-usage, if anyone is
willing to discuss that with me.
More information about the Maintenance