Re: [monit] monit fails mysteriously on refresh after upgrade from Debian etch to lenny

[Jan-Henrik Haukeland]

Doesn't sound like Monit crash, it just runs one cycle and then dies.  
The obvious explanation is that Monit was not started in daemon mode  
and that this is a configure issue.

So can you verify that you have a "set daemon x" in the monitrc file,  
where the x represents a number?  Also that you use the receipt for  
running Monit from init described here, http://mmonit.com/wiki/Monit/FAQ#init

If yes, could you please provide us with the output from Monit's log  
file when you run Monit in debug mode (the -v switch to the program)  
and your monitrc file?

On Feb 23, 2009, at 4:32 PM, Jenny Hopkins wrote:

> After upgrading from etch to lenny, monit won't run for more than  
> one cycle.
> It is started from /etc/init.d - starts fine, runs through all its
> tests, the process sleeps.
> When it reaches time to check again, it just dies.  I've changed
> timings from wait 180 secs to wait 10 secs/20 secs etc, and always the
> same.
>
> It's not likely the monitrc file, because I've tried using the setup
> with the empty file supplied and only the very basics put in such as
> server and logfile - no checking to do.
>
> I tried running an strace on it, and this gives an enormous output,
> but I don't know what to look for.
> Also the trace runs while monit is still up.
> Something like this line could be a clue, but could also be a red  
> herring:
> access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or  
> directory)
>
> I'd have thought if it were a library problem monit wouldn't even
> start in the first place.
>
> Can anyone give me a hint as to what exactly happens when monit
> refreshes, rather than starts?
>
> I'm at the hair-tearing out stage....
>
> Files can be supplied.  Monit version is 1:4.10.1-4.
> Thanks,
>
> Jenny