www-apache-bugdb mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ratman <...@rat.org>
Subject Re: general/1724: endless copies of the deamon spawn, making server not respond
Date Mon, 26 Jan 1998 20:18:34 GMT
On 23 Jan 1998 dgaudet@hyperreal.org wrote:

> [In order for any reply to be added to the PR database, ]
> [you need to include <apbugs@Apache.Org> in the Cc line ]
> [and leave the subject line UNCHANGED.  This is not done]
> [automatically because of the potential for mail loops. ]
> Synopsis: endless copies of the deamon spawn, making server not respond
> State-Changed-From-To: open-analyzed
> State-Changed-By: dgaudet
> State-Changed-When: Fri Jan 23 13:56:26 PST 1998
> State-Changed-Why:
> Even if it isn't NFS it's still likely a local problem... I
> maintain a number of linux boxes running apache without this
> happening.  Are you absolutely sure that your load just isn't
> high enough to cause that?  Like maybe a CGI is being hit a
> whole bunch and chewing up all the clients.  You should look
> at the error_log, /server-status, and even the access_log to
> see just what is being requested.
> Dean

It hits the 150 child limit, so it doesn't take down the system or even
slow it down that bad. Load average is still under 2. And no CGIs would be
running at that time at all, so I don't think that's it. Unless they're
exiting, and somehow leaving the child process hosed.
It just makes me wonder, cause this doesn't happen with the old version.
I've been compiling different apache versions on this machine for a long
time, and I'm not doing anything different than usual.

Oh, and it will say "httpd <zombie>" or something like that next to a lot 
of them, does that help?

I've been running the old one to see if that fixes it and it does. So it's
definatly something about the new version that's triggering the problem.
Yesterday I put on the newest release kernel (2.0.33) to see if that helps
and put back the newer apache. I'll gather some data (i'll do a ps -alx)
next time it's clogged with dead children (boy does that sound sick). I
can use the system fine, again, it's hitting that 150 child limit.
God, i hope it's nothing on my end. I'd hate to be wasting your time with
a support issue. :(

View raw message