Received: by taz.hyperreal.com (8.6.12/8.6.5) id EAA23552; Fri, 19 Apr 1996 04:44:40 -0700 Received: from arachnet.algroup.co.uk by taz.hyperreal.com (8.6.12/8.6.5) with SMTP id EAA23537; Fri, 19 Apr 1996 04:44:29 -0700 Received: from heap.ben.algroup.co.uk by arachnet.algroup.co.uk id aa19339; 19 Apr 96 12:43 BST Received: from gonzo.ben.algroup.co.uk by heap.ben.algroup.co.uk id aa11900; 19 Apr 96 12:12 BST Subject: Re: socket failed To: new-httpd@hyperreal.com Date: Fri, 19 Apr 1996 12:07:32 +0100 (BST) From: Ben Laurie In-Reply-To: <199604191043.FAA05181@caffeine.io.com> from "James H. Cloos Jr." at Apr 19, 96 05:43:17 am X-Mailer: ELM [version 2.4 PL24 PGP2] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 1686 Message-ID: <9604191207.aa06215@gonzo.ben.algroup.co.uk> Sender: owner-new-httpd@apache.org Precedence: bulk Reply-To: new-httpd@hyperreal.com Status: O X-Status: James H. Cloos Jr. wrote: > > I also get tons of socket error: accept failed errors. The lastest > snapshots change the error from something like: > > [Fri Apr 12 12:20:46 1996] socket error: accept failed > > to something like: > > [Thu Apr 18 22:29:15 1996] accept: Connection reset by peer > [Thu Apr 18 22:29:15 1996] - socket error: accept failed > > We got just under 7000 of those on the 18th, and just under 375000 > lines in the access_logs. > > The ratio was higher earlier in the week, but has settled back down to > the normal 1 to 2 percent. That would be consistent with the state of the net earlier in the week. It would seem that this error is simply a symptom of impatient users dropping the connection. Of course, it may be a symptom of frustrated users waiting for a long time and then dropping the connection. Presumably what happens is that during the initial handshake some packets go missing and when they finally get through the client has already closed the socket. It would be really nice to see a packet trace of one of these events to confirm this theory ... anyone kitted up to do that? Cheers, Ben. > > The server and kernel are both kept nearly current on that box. (The > 2nd half of the 18th I ran 1.3.90 and the 19960418130013 CVS snapshot, > the first half had a 4 or 5 day old.) > > -JimC > -- > James H. Cloos, Jr. > cloos@io.com LPF,Usenix,SAGE,ISOC,ACLU -- Ben Laurie Phone: +44 (181) 994 6435 Freelance Consultant and Fax: +44 (181) 994 6472 Technical Director Email: ben@algroup.co.uk A.L. Digital Ltd, URL: http://www.algroup.co.uk London, England.