non-blocking test scheduler (faulty service slows down monit)

Issue #562 open
ethaniel 1 created an issue

I have a service (sphinx) which on start immediately gives a "permission denied error". For some reason it's blocking monit for 30 seconds every time it is started, even though when I start it by hand it returns the error instantly.

I am using monit 5.20.0

[MSK Feb 24 11:48:32] error    : 'sphinx' process is not running
[MSK Feb 24 11:48:32] info     : 'sphinx' trying to restart
[MSK Feb 24 11:48:32] info     : 'sphinx' start: '/bin/bash -c /etc/init.d/searchd start'
[MSK Feb 24 11:49:02] error    : 'sphinx' failed to start (exit status 1) -- '/bin/bash -c /etc/init.d/searchd start': Starting searchd: [Fri Feb 24 11:48:32.707 2017] [22609] using config file '/etc/sphinx/sphinx.conf'...
[Fri Feb 24 11:48:32.707 2017] [22609] FATAL: failed to open log file 'searchd.log': Permission denied
Sphinx
[MSK Feb 24 11:49:02] error    : 'proxysql_monitor' process is not running

Comments (9)

  1. ethaniel 1 reporter

    What I don't understand is why doesn't monit check other processes in those 30 seconds? The error is returned instantly, I would assume that a faulty program should not halt monit for 30 seconds.

  2. Tildeslash repo owner

    Monit currently performs the checks serially and waits for up to timeout seconds for the process to start.

    We will implement new test scheduler in Monit, which will provide non-blocking tests and will allow to run tests in parallel.

  3. Log in to comment