From owner-cvs-all Tue Feb 19 15:39: 7 2002 Delivered-To: cvs-all@freebsd.org Received: from patrocles.silby.com (d23.as29.nwbl0.wi.voyager.net [169.207.73.23]) by hub.freebsd.org (Postfix) with ESMTP id F0E1E37B404; Tue, 19 Feb 2002 15:38:59 -0800 (PST) Received: from patrocles.silby.com (localhost [127.0.0.1]) by patrocles.silby.com (8.12.2/8.12.2) with ESMTP id g1JHgiFY004528; Tue, 19 Feb 2002 17:42:44 GMT (envelope-from silby@silby.com) Received: from localhost (silby@localhost) by patrocles.silby.com (8.12.2/8.12.2/Submit) with ESMTP id g1JHgf7U004525; Tue, 19 Feb 2002 17:42:43 GMT X-Authentication-Warning: patrocles.silby.com: silby owned process doing -bs Date: Tue, 19 Feb 2002 17:42:41 +0000 (GMT) From: Mike Silbersack To: Sheldon Hearn Cc: cvs-committers@FreeBSD.org, Subject: Re: cvs commit: src/sys/kern kern_fork.c In-Reply-To: <15675.1014160079@axl.seasidesoftware.co.za> Message-ID: <20020219173356.S3645-100000@patrocles.silby.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG On Wed, 20 Feb 2002, Sheldon Hearn wrote: > On Mon, 18 Feb 2002 19:15:28 PST, Mike Silbersack wrote: > > > Log: > > A few misc forkbomb defenses: > > > > - Remove the printing of "proc: table full". When the table > > really is full, this would flood the screen/logs, making > > the problem tougher to deal with. > > I like the idea that my console won't be filled with crap, but what clue > will I get now that kern.maxfiles needs to be bumped? > > Ciao, > Sheldon. This affects maxproc, not maxfiles. If you are hitting the limit, I expect that your apps would start complaining. If ratelimited, those message could certainly be useful. If I get some time, I'll look into doing that. Actually, what would be a good idea is to have one kernel thread that woke up every few seconds and reported resource shortages, then the various subsystems which have all these warning messages could just export their counters, and the "warning daemon" could print the warnings occasionally. (The people could get fancy and add pre-failure diagnostics like "you've been running at 75% maxproc usage, it's time to up maxusers or add more ram.") But in the meantime, program logs or top would be the best way to monitor for that condition. The non-ratelimited messages were just too much of a DoS themselves. Mike "Silby" Silbersack To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message