www-apache-bugdb mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dean Gaudet <dgau...@arctic.org>
Subject Re: FW: Re: general/1540: Apache stops to respond on virtual interfa
Date Fri, 12 Dec 1997 18:50:00 GMT
The following reply was made to PR general/1540; it has been noted by GNATS.

From: Dean Gaudet <dgaudet@arctic.org>
To: Lars Eilebrecht <Lars.Eilebrecht@unix-ag.org>
Cc: apbugs@apache.org, Andreas Norman <andreas@dalnet.se>
Subject: Re: FW: Re: general/1540: Apache stops to respond on virtual interfa
Date: Fri, 12 Dec 1997 10:52:50 -0800 (PST)

 Is NFS involved in any way?  If yes, then try using the LockFile directive
 like "LockFile /var/tmp/htlock" to move it away from NFS.
 Does the /server-status page show there are idle children?  If not,
 what are they doing?  If they're stuck in DNS lookup then either disable
 HostnameLookups, or turn off the solaris name caching daemon... whose
 name escapes me at the moment.
 Are you using multiple Listen statements?
 Can you reproduce this on a server which is running on a different port,
 and using the -X option?  This will cause it to run only a single child.
 If you can't reproduce it with -X then try setting MaxSpareServers
 to 2, MinSpareServers to 1, and see if you can reproduce it that way.
 Keep trying to increase MaxSpare until you can reproduce it.
 Once you've got it to reproduce like this, use truss to find out what
 system call each child is doing.  If they're all idle, then all but one
 of them should be in fcntl(), the rest should be in select().

View raw message