Mongrel users - Jan 2008 - writing pid file earlier

I''m using god* to monitor my mongrels. God looks at pid files to know
the status of the cluster. It seems as though mongrel does not write a
pid file until it has loaded the entire rails/project environment,
which in my case, takes upwards of 90 seconds. Meanwhile, god is
freaking out because it thinks that there aren''t mongrels running, and
it tries to start them. Thankfully I can set god''s "grace
period" to 5
minutes, so I only get one false restart, but it would be nice to have
a different solution.

Short of patching god the only thing I can think of is some sort of
wrapper shell or ruby script that writes the pid ahead of time, but
even then, the pid would conflict with when mongrel wants to write a
pid-- I can''t think of a good way to delete the initial pid "just
in
time".

Any ideas are appreciated, or if the mongrel team is open to a patch
to write the pid much earlier, I can try to put that together.

Cheers,
John

* http://god.rubyforge.org/

-- 
John Joseph Bachir
http://blog.johnjosephbachir.org
http://lyceum.ibiblio.org
http://dissent.cc
http://jjb.cc

Lets say you alter the code.

What happens if the PID file is written, but Mongrel dies kicking off the Rails
app (because you have a problem with your Rails code)? Won''t god* then
think that your process is running when it isn''t? Isn''t this
false positiver worse than a couple false negatives?

Seems to me that if you can''t tell god* to wait long enough for your
process to start, maybe god* isn''t the answer...

=Will Green

John Joseph Bachir wrote ..
> I''m using god* to monitor my mongrels. God looks at pid files to
know
> the status of the cluster. It seems as though mongrel does not write a
> pid file until it has loaded the entire rails/project environment,
> which in my case, takes upwards of 90 seconds. Meanwhile, god is
> freaking out because it thinks that there aren''t mongrels running,
and
> it tries to start them. Thankfully I can set god''s "grace
period" to 5
> minutes, so I only get one false restart, but it would be nice to have
> a different solution.
> 
> Short of patching god the only thing I can think of is some sort of
> wrapper shell or ruby script that writes the pid ahead of time, but
> even then, the pid would conflict with when mongrel wants to write a
> pid-- I can''t think of a good way to delete the initial pid
"just in
> time".
> 
> Any ideas are appreciated, or if the mongrel team is open to a patch
> to write the pid much earlier, I can try to put that together.
> 
> Cheers,
> John
> 
> * http://god.rubyforge.org/
> 
> -- 
> John Joseph Bachir
> http://blog.johnjosephbachir.org
> http://lyceum.ibiblio.org
> http://dissent.cc
> http://jjb.cc
> _______________________________________________
> Mongrel-users mailing list
> Mongrel-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/mongrel-users
>

On Jan 24, 2008 11:43 AM, Will Green <will at hotgazpacho.com>
wrote:
> Lets say you alter the code.
>
> What happens if the PID file is written, but Mongrel dies kicking off the
Rails app (because you have a problem with your Rails code)? Won''t god*
then think that your process is running when it isn''t? Isn''t
this false positiver worse than a couple false negatives?
Well, IMHO, mongrel should write the pid file early, and clean the pid
file up when it dies, for whatever reason short of something
untrappable like a segfault.


Kirk Haines