Received: by taz.hyperreal.com (8.6.12/8.6.5) id UAA08499; Thu, 21 Dec 1995 20:31:41 -0800 Received: from life.ai.mit.edu by taz.hyperreal.com (8.6.12/8.6.5) with SMTP id UAA08494; Thu, 21 Dec 1995 20:31:38 -0800 Received: from volterra.ai.mit.edu by life.ai.mit.edu (4.1/AI-4.10) for new-httpd@hyperreal.com id AA23492; Mon, 18 Dec 95 14:05:57 EST From: rst@ai.mit.edu (Robert S. Thau) Received: by volterra.ai.mit.edu (8.6.12/AI-4.10) id OAA25397; Mon, 18 Dec 1995 14:05:54 -0500 Date: Mon, 18 Dec 1995 14:05:54 -0500 Message-Id: <199512181905.OAA25397@volterra.ai.mit.edu> To: new-httpd@hyperreal.com Subject: Re: Number of Virtual Servers Sender: owner-new-httpd@apache.org Precedence: bulk Reply-To: new-httpd@apache.org Given that any server that would run into these problems has to be admined by someone with a clue, we could get around this by suggesting the server run in a group that has write access to the logs. ... thus, in effect, giving the server (and, implicitly, potential rogue CGI scripts) permission to open the logs as well. (This is why permission to open the log files is an issue). I'm not in the position of trying to run with large numbers of virtual hosts here, so my comments may be a bit off base. Still, I haven't seen much reaction to the suggestion of having a single log file with entries labeled with the vhost, for later splitting if necessary; this neatly handles the descriptor hogging problem. Any comments? rst