Return-Path: Delivered-To: new-httpd-archive@hyperreal.org Received: (qmail 22930 invoked by uid 6000); 4 Dec 1998 23:14:42 -0000 Received: (qmail 22921 invoked from network); 4 Dec 1998 23:14:41 -0000 Received: from twinlark.arctic.org (204.62.130.91) by taz.hyperreal.org with SMTP; 4 Dec 1998 23:14:41 -0000 Received: (qmail 19025 invoked by uid 500); 4 Dec 1998 23:14:35 -0000 Date: Fri, 4 Dec 1998 15:14:35 -0800 (PST) From: Dean Gaudet To: new-httpd@apache.org Subject: Re: Status bug FIX ?? In-Reply-To: <199812042113.QAA17232@devsys.jaguNET.com> Message-ID: X-Comment: Visit http://www.arctic.org/~dgaudet/legal for information regarding copyright and disclaimer. MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: new-httpd-owner@apache.org Precedence: bulk Reply-To: new-httpd@apache.org On Fri, 4 Dec 1998, Jim Jagielski wrote: > Dean Gaudet wrote: > > > > On Fri, 4 Dec 1998, Harrie Hazewinkel wrote: > > > > > And since we only use the "server_hostname" of the server_rec > > > for "mod_status" I now use the "server_hostname" ptr directly > > > in the scoreboard. > > > > That, too, breaks across restart. > > > > Jim committed a fix which works in most cases, but still has a race > > condition. The race condition is when a child is in the middle of serving > > a status request when a graceful restart happens, it will start reading > > server_rec's from children of the newer generation. > > > > Is that bad, ie: reading server_rec's from the _newer_ generation? At > least it's valid data. I had thought that the only problem was reading > server_rec info from the _previous_ generation, which would be bogus. > I admit that I hadn't considered the former... Yup it's bad to read any pointer from another generation, newer or older, because the parent used to spawn the other child didn't have the same memory map. Dean