On Tue, 30 Sep 1997 [EMAIL PROTECTED] wrote: > Date: Tue, 30 Sep 1997 04:48:45 -0700 (PDT) > From: [EMAIL PROTECTED] > To: [EMAIL PROTECTED], [EMAIL PROTECTED], [EMAIL PROTECTED] > Subject: Re: general/1181: Too many virtual hosts cause cgi scripts to stop > working > > Synopsis: Too many virtual hosts cause cgi scripts to stop working > > State-Changed-From-To: open-closed > State-Changed-By: [EMAIL PROTECTED] > State-Changed-When: Tue Sep 30 04:48:44 PDT 1997 > State-Changed-Why: > > This a frequently asked question. > See http://www.apache.org/docs/misc/FAQ.html#fdlim > for details.
Yes, I did find that just after submitting the bug. This is why I'm glad I marked the bug 'Request for Improvement'. I really think you should get it to print a useful error message to the logs, other than 'could not spawn child'. Possibly have the configure script determine the hard file descriptor limit, or when the 'could not spawn child' error message is about to be printed, try to open /dev/null two or three times and check for errno==EMFILE; if so, change the error message to say 'out of file descriptors'. The 'silent failure' and/or 'useless error message' is something I'd expect from Microsoft, not Apache! Now that I've said my piece, I have to ask you how to set up two separate httpd processes, one handling the first 75 ip-based virtual hosts and the second handling the last 75. I tried putting a Listen directive in each VirtualHost, but that is disallowed. What is the right method for splitting the job like this? (I can't believe that I might have to start 150 different httpd processes. That'd be stupid.) > Severity-Changed-From-To: serious-non-critical > Severity-Changed-By: [EMAIL PROTECTED] > Severity-Changed-When: Tue Sep 30 04:48:44 PDT 1997 -- Charles Howes -- [EMAIL PROTECTED]
