Received: by taz.hyperreal.com (8.8.4/V2.0) id QAA10181; Tue, 25 Feb 1997 16:31:54 -0800 (PST) Received: from sierra.zyzzyva.com by taz.hyperreal.com (8.8.4/V2.0) with ESMTP id QAA10067; Tue, 25 Feb 1997 16:31:39 -0800 (PST) Received: from scanner.worldgate.com (scanner.worldgate.com [198.161.84.3]) by sierra.zyzzyva.com (8.8.4/8.8.2) with ESMTP id SAA00221 for ; Tue, 25 Feb 1997 18:20:03 -0600 (CST) Received: from znep.com (uucp@localhost) by scanner.worldgate.com (8.8.5/8.7.3) with UUCP id RAA16873 for new-httpd@hyperreal.com; Tue, 25 Feb 1997 17:20:18 -0700 (MST) Received: from localhost (marcs@localhost) by alive.znep.com (8.7.5/8.7.3) with SMTP id RAA24071 for ; Tue, 25 Feb 1997 17:12:53 -0700 (MST) Date: Tue, 25 Feb 1997 17:12:53 -0700 (MST) From: Marc Slemko To: Apache Group Subject: Re: [BUG]: "HTML files crate zombies " on FreeBSDLinux (fwd) In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: new-httpd-owner@apache.org Precedence: bulk Reply-To: new-httpd@hyperreal.com This could be related to the problem with the server having to kill child processes that has been reported. But probably not if it happens with 1.1, but I wouldn't trust them to know what version they are using... Yes, the reporter has no clue about how the heck anything works, but... On Tue, 25 Feb 1997, Rob Hartill wrote: > > This seems to be complete nonsense unless I'm misreading it. > > not acked. > > ---------- Forwarded message ---------- > Date: Mon Feb 24 23:46:26 1997 > From: sk@www.russia.net > To: apache-bugs%apache.org@organic.com > Subject: [BUG]: "HTML files crate zombies " on FreeBSDLinux > > Submitter: sk@www.russia.net > Operating system: FreeBSDLinux, version: > Version of Apache Used: 1.1.1, 1.2.something > Extra Modules used: no dependance > URL exhibiting problem: > > Symptoms: > -- > The html : > > #! /usr/bin/perl > print "Content-type: text/html\n\n"; > print "\n"; > $cnt=0; > $maxcnt=10; > while($cnt!=$maxcnt){ > print "\n"; > $cnt++; > } > > print "\n"; > > > This is just enough to start zombie which will be > the httpd responsible for. > > Zombies will be the 'somebinary's > > The number of zombie creation might range in 1,2,... > (not exactly one zombie) > > the frequency of zombie emergence will greately depend on the > system load average, and in order to watch you may have to > sit waiting for a while. > > The problem reported by the hackers is the > incorrect loop structure in the alloc.c > fork()/signal()/wait() handling. > > The potential problems arizing seem to be easy to > model by waiting in the perl script which produces the HTML > page, while allocing lots of memory in the child 'somebinary' > > Regards, > Serguei Kouzmine > > -- > > Backtrace: > -- > > -- > >