Re: monit_http

[Jan-Henrik Haukeland]

Rory Toma <address@hidden> writes:


> I'll have to poke in the kernel to see what's going on.

in the tcp/ip stack implementation. And again, Google for "socket and
TIME_WAIT" since it's in this problem domain the problem lies, Stevens
does also discuss the TIME_WAIT problem in his "Unix Network
programming" book.


> > BTW, how long do you have to wait before monit can bind the server
> > socket again after you do a skill -9 monit?  If the wait is in
> > seconds, the best thing to do, I agree, is to just exit, but if
> > it's short it may be worth looping a few times in
> 
> It's from a fraction of a second to seconds. I'll check in the exit
> if Run.init and I'll see if I can find the root cause of the problem
> before doing more.

I'm not sure you will find an easy solution to the problem :-( 

BTW, before check in; maybe use kill_daemon(SIGTERM) instead of exit()
so the daemon is taken down gracefully and with a log message?

-- 
Jan-Henrik Haukeland