httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dean Gaudet <dgau...@arctic.org>
Subject Re: did someone play with child spawn/kill code?
Date Mon, 12 Oct 1998 14:04:12 GMT
I don't see anything immediately obvious in the 1.3.2 -> 1.3.3 diffs or in
the ssl patches...

When I straced the child the parent was trying to kill, it was in flock() 
and had SIGUSR1 blocked.  "That shouldn't happen".  Although there is a
certain amount of arbitrary code between the enabling of the SIGUSR1
handler and the accept_mutex_on in the main loop -- because the pools are
cleared after enabling SIGUSR1. 

My server has php3 installed... I noticed that php blocks SIGUSR1 in
functions/hg_comm.c, dunno what that means though.  Rasmus, is it possible
that this function would be called from a registered cleanup?

Dean

On Mon, 12 Oct 1998, Dean Gaudet wrote:

> My server has all sorts of this crap:
> 
> [Mon Oct 12 05:57:28 1998] [info] server has 49 children, 49 are idle, killing one
> [Mon Oct 12 05:57:29 1998] [info] server has 49 children, 49 are idle, killing one
> [Mon Oct 12 05:57:30 1998] [info] server has 49 children, 49 are idle, killing one
> [Mon Oct 12 05:57:31 1998] [info] server has 49 children, 49 are idle, killing one
> [Mon Oct 12 05:57:32 1998] [info] server has 49 children, 49 are idle, killing one
> [Mon Oct 12 05:57:33 1998] [info] server has 49 children, 49 are idle, killing one
> [Mon Oct 12 05:57:34 1998] [info] server has 49 children, 49 are idle, killing one
> [Mon Oct 12 05:57:35 1998] [info] server has 49 children, 49 are idle, killing one
> [Mon Oct 12 05:57:36 1998] [info] server has 49 children, 49 are idle, killing one
> [Mon Oct 12 05:57:37 1998] [info] server has 49 children, 49 are idle, killing one
> [Mon Oct 12 05:57:38 1998] [info] server has 49 children, 49 are idle, killing one
> [Mon Oct 12 05:57:39 1998] [info] server has 49 children, 49 are idle, killing one
> [Mon Oct 12 05:57:40 1998] [info] server has 49 children, 49 are idle, killing one
> [Mon Oct 12 05:57:41 1998] [info] server has 49 children, 49 are idle, killing one
> [Mon Oct 12 05:57:42 1998] [info] server has 49 children, 48 are idle, killing one
> [Mon Oct 12 05:57:43 1998] [info] server has 49 children, 48 are idle, killing one
> [Mon Oct 12 05:57:44 1998] [info] server has 49 children, 48 are idle, killing one
> 
> which is a message I don't think is in the stock code... anyhow, this
> wasn't happenning with code from the beginning of sept, and it's now
> happenning with code from the end of sept.  Although it could be the SSL
> patches... or something else.  I haven't had a chance to look at it yet.
> It tries to kill a child every second -- it's not spawning children, that
> would be logged too.
> 
> Dean
> 
> 


Mime
View raw message