The following reply was made to PR general/1181; it has been noted by GNATS.
From: Charles Howes <[EMAIL PROTECTED]> To: Dean Gaudet <[EMAIL PROTECTED]> Subject: Re: general/1181: Too many virtual hosts cause cgi scripts to stop working Date: Tue, 30 Sep 1997 16:06:58 -0700 (PDT) On Tue, 30 Sep 1997, Dean Gaudet wrote: > Return-Path: <[EMAIL PROTECTED]> > Received: from swiss.ics.bc.ca (swiss3.ics.bc.ca [204.244.222.249]) by > cheddar.ics.bc.ca (8.8.5/8.8.3) with ESMTP id PAA21462 for <[EMAIL > PROTECTED]>; Tue, 30 Sep 1997 15:30:11 -0700 (PDT) > Received: from twinlark.arctic.org (twinlark.arctic.org [204.62.130.91]) by > swiss.ics.bc.ca (8.8.5/8.8.5) with SMTP id PAA12216 for <[EMAIL PROTECTED]>; > Tue, 30 Sep 1997 15:32:05 -0700 (PDT) > Received: (qmail 21757 invoked by uid 500); 30 Sep 1997 22:31:18 -0000 > Date: Tue, 30 Sep 1997 15:31:18 -0700 (PDT) > From: Dean Gaudet <[EMAIL PROTECTED]> > To: Charles Howes <[EMAIL PROTECTED]> > cc: [EMAIL PROTECTED], [EMAIL PROTECTED] > Subject: Re: general/1181: Too many virtual hosts cause cgi scripts to stop > working > In-Reply-To: <[EMAIL PROTECTED]> > Message-ID: <[EMAIL PROTECTED]> > MIME-Version: 1.0 > Content-Type: TEXT/PLAIN; charset=US-ASCII > > The error is detected in the child, after the fork, at which time there is > no clear communication channel back to the parent. In fact it does log > the errno from exec. What you suggest is not portable if it masks another > error from exec, and how are we to know, without reading all the man pages > and trying it out everywhere, that some OS just happens to do things > correctly? Sigh. It all seemed so easy. > On Tue, 30 Sep 1997, Charles Howes wrote: > > > (And it'll save countless web admins headaches as they try and find > > the cause of the mysterious failure of all cgi-bins on the system > > simultaneously...) > > But the reason for this is almost always because they have far too may > file descriptors open. For which we have another solution: don't open up > an access log for each vhost. In 1.3 you can feed a pipe reliably with > your log, and the logger can split the log in real time. Hmm... guess I should change from 1.2.4 to 1.3... :-) -- Charles Howes -- [EMAIL PROTECTED]
