Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 25 Mar 2001 18:05:48 -0600 (CST)
From:      Ryan Thompson <ryan@sasknow.com>
To:        "Eric D. Stanfield" <exs@kka.com>
Cc:        freebsd2 <freebsd-isp@FreeBSD.ORG>
Subject:   Re: server problem
Message-ID:  <Pine.BSF.4.21.0103251802130.66246-100000@ren.sasknow.com>
In-Reply-To: <003001c0b47f$485a5620$7ccc29d0@thestanfields.com>

next in thread | previous in thread | raw e-mail | index | archive | help

I don't know off hand what might be causing this.

Random thought: If the server process is a daemon, it will fork and return
control. If it starts successfully, but for some reason returns a non-zero
exit status, your start/stop script might be fooled into thinking that it
did not start, thus it tries to start it again, etc., ad infinitum. Check
your logic. Maybe make a fancy grep command to ensure that no processes by
the same name are already running. Or better yet, put a .pid file in
/var/run/ (if it doesn't do that already).

Other than that, assuming that you eventually got those processes killed,
check your logs and analyze what you can. In particular, have a look at
the times the server was started--in very close succession, or at specific
intervals?

That should also tell you if this problem took a few days to cook up, and
the servers simply weren't dying as expected at 8am.




Eric D. Stanfield wrote to freebsd2:

> Wondering if anyone else has ever had this happen.
> 
> I have a server set up that runs the online game CounterStrike.  Basically,
> cron launches counterstrike every evening at 5pm then turns it off at 8am by
> calling a little shell script (for both launch and kill).  Counterstrike
> sets aside 128M of ram for its own use.  This has worked well for months.
> 
> Yesterday, cron turned counterstrike on.  Then cron did it again.  And
> again.  I telnetted in from home and saw around 12-15 counterstrike
> processes fighting to load/stay loaded.  The login process itself took about
> 20 minutes to get from prompting for a username to giving me a command
> prompt.  I barely managed to do a ps -ax to see all the c-s processes before
> getting dumped out of telnet.  At that point it would appear that inetd died
> because ftp/telnet no longer accept connections.  Oddly enough, Apache is
> still running and I can pull a web page off the machine albeit at a snail's
> pace.
> 
> As I said, the server has been running this scenario without problems for
> months.  I've not edited cron or changed anything else on the machine.  The
> server is running 4.2-Current and is a p2-266 with 256M of ram and endless
> gigs of disk space.
> 
> Has anyone seen this happen before?  I'm guessing that cron is bugging out
> and repeatedly launching this counterstrike process.  I can't find anything
> related to bugs with cron in this release.  While this is "only a game
> machine"  I do run a number of production machines on 4.2-current and I'm
> very worried at this point that I might find myself dealing with this in a
> much more critical environment.
> 
> Thanks,
> 
> Eric
> 
> 
> To Unsubscribe: send mail to majordomo@FreeBSD.org
> with "unsubscribe freebsd-isp" in the body of the message
> 

-- 
  Ryan Thompson <ryan@sasknow.com>
  Network Administrator, Accounts

  SaskNow Technologies - http://www.sasknow.com
  #106-380 3120 8th St E - Saskatoon, SK - S7H 0W2

        Tel: 306-664-3600   Fax: 306-664-1161   Saskatoon
  Toll-Free: 877-727-5669     (877-SASKNOW)     North America


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-isp" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.21.0103251802130.66246-100000>