Received: (from majordom@localhost)
	by hyperreal.org (8.8.5/8.8.5) id MAA03643;
	Tue, 22 Jul 1997 12:20:20 -0700 (PDT)
Received: (from gnats@localhost)
	by hyperreal.org (8.8.5/8.8.5) id MAA03464;
	Tue, 22 Jul 1997 12:20:04 -0700 (PDT)
Date: Tue, 22 Jul 1997 12:20:04 -0700 (PDT)
Message-Id: <199707221920.MAA03464@hyperreal.org>
From: Gerald Anderson <gander@netcomi.com>
Reply-To: Gerald Anderson <gander@netcomi.com>
To: apache-bugdb@apache.org
Cc: apache-bugdb@apache.org
Subject: config/897: Possible Starving Sockets issue.  In same conf 1 virtual
 getting connection regused, others ok. Intermittant
In-Reply-To: Your message of Tue, 22 Jul 1997 12:19:56 -0700 (PDT)
	<199707221919.MAA03363@hyperreal.org>
Sender: apache-bugdb-owner@apache.org
Precedence: bulk


>Number:         897
>Category:       config
>Synopsis:       Possible Starving Sockets issue.  In same conf 1 virtual getting connection regused, others ok. Intermittant
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    apache (Apache HTTP Project)
>State:          open
>Class:          sw-bug
>Submitter-Id:   apache
>Arrival-Date:   Tue Jul 22 12:20:02 1997
>Originator:     gander@netcomi.com
>Organization:
apache
>Release:        1.2.1
>Environment:
Linux multi1.netcomi.com 2.0.29 #7 Fri Jun 20 08:22:21 CDT 1997 i686
gcc version 2.7.2.1> httpd -v 
Server version Apache/1.2.1.

>Description:
We run 8 copies of httpd with 8 different config files using listens and virtual
hosts.  Recently (since 1.2.1 upgrade?) It seems as though we are seeing a starving socket
type problem where one virtual will get connection refused, and another is fine. . .same
config.  It seems pretty intermittent, and we haven't found a reliable way
of reproducing the problem.  We do see it fairly often with nearly 14,000 virtual
web sites (on 35 servers).  All the machines are fairly busy, but have been for
months.
>How-To-Repeat:
Unknown
>Fix:
Verify that it is/isn't starving sockets on linux (linux does serial_accept by default).
Obvously no serial_accept compiles fix this issue. 
>Audit-Trail:
>Unformatted: