The following reply was made to PR general/3312; it has been noted by GNATS.

From: Ole Tange <[EMAIL PROTECTED]>
To: [EMAIL PROTECTED], [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]
Subject: Re: general/3312: Children die. Parent stops serving requests
Date: Mon, 2 Nov 1998 00:13:39 +0100 (CET)

 On Fri, 30 Oct 1998, Ole Tange wrote:
 
 > As I see the problem:
 > 
 > The parent thinks it has a lot of children that are terrible slow at
 > processing requests (SS is rather high if you look at 'apachectl status').
 > So it spawns new children until it reaches the MaxClient-limit
 > 
 > But what (I think) really happens is that the children die without
 > notifying the parent.
 
 Assuming my theory is correct, then I ought to get a much longer uptime if
 I raise MaxClient.
 
 I have now raised MaxClient to 256. I'll let you know if it improved the
 uptime. I *do* know that I am playing with fire: If I get 256 actual
 clients the server will probably swap until death.
 
 To the apache-team: Assuming my theory is correct, then the problem would
 be solved (or at least: worked around) if apache ran a 'ps -aux' (or
 equivalent) once in a while to see if some of the children died without
 telling.
 
 
 /Ole
 
 

Reply via email to