From owner-freebsd-questions  Sun Mar 16 14:51:26 1997
Return-Path: <owner-questions>
Received: (from root@localhost)
          by freefall.freebsd.org (8.8.5/8.8.5) id OAA23784
          for questions-outgoing; Sun, 16 Mar 1997 14:51:26 -0800 (PST)
Received: from caliban.dihelix.com (caliban.mrtc.org [199.4.33.251])
          by freefall.freebsd.org (8.8.5/8.8.5) with ESMTP id OAA23760;
          Sun, 16 Mar 1997 14:51:14 -0800 (PST)
Received: (from langfod@localhost) by caliban.dihelix.com (8.8.4/8.8.3) id MAA01538; Sun, 16 Mar 1997 12:51:05 -1000 (HST)
Message-Id: <199703162251.MAA01538@caliban.dihelix.com>
Subject: Re: problem with over 133 virutal domains with apache on 2.1.7
In-Reply-To: <Pine.BSF.3.95.970316155552.6217E-100000@alpha.risc.org> from Brian Tao at "Mar 16, 97 03:58:32 pm"
To: taob@risc.org (Brian Tao)
Date: Sun, 16 Mar 1997 12:51:04 -1000 (HST)
Cc: langfod@dihelix.com, questions@FreeBSD.org, isp@FreeBSD.org
From: "David Langford" <langfod@dihelix.com>
X-blank-line: This space intentionaly left blank.
X-Mailer: ELM [version 2.4ME+ PL31 (25)]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-questions@FreeBSD.org
X-Loop: FreeBSD.org
Precedence: bulk

Brian Tao
>On Wed, 12 Mar 1997, David Langford wrote:
>> 
>> When I add my 134'th virtual domain to apache it says it can no
>> longer figure out its hostname and wants me to set ServerName
>
>    You are probably running out of file descriptors (defaults to
>FD_SETSIZE=256 per process) because you have separate access and error
>logs for each virtual domain.  You'll probably have to rebuild your
>kernel with FD_SETSIZE=1024 (or whatever) and recompile Apache with
>that value as well.
>
>    Also, are you running named on that box?  I think named by default
>Brian Tao (BT300, taob@risc.org)

What I did to get it working was add -DFD_SETSIZE=1024 to make.conf
Rebuilt kernel/libc and apache.  Everything works okay but the
system seem to be unstable- reboots about once a day (of course
only when I am not around).

Strangly enough named on this machine has NOT been a problem *shrug*.
Also gated pukes after a bit with assert failed in rt_radix.c
assert(RT_TABLE_PTR(socktype(rth->rth_dest)

netstat -m shows 503/704 mbub clusters about 3hours after last reboot
on a Sunday.

My main apache proccess show 336 open files and 44781 total at the moment.
(fstat of root http proccess and fstat of web user)

I think I am going to redo this with FD_SETSIZE=512 and see if that helps the 
system.

I wish I knew why the machine was crashing right now.  I can only tell
if I am on the console when it does and I dont want to set gdb as
I prefer it to come back alive quickly.

Any thoughts if 2.2 has any tweaks to handle this better. I have
forgotten the changes fyling by in the past few months :) 

Thanks to everybody who has sent help on this. :)

-David Langford
 langfod@dihelix.com