From owner-freebsd-stable Thu Jul 11 17:57:45 1996 Return-Path: owner-stable Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id RAA17664 for stable-outgoing; Thu, 11 Jul 1996 17:57:45 -0700 (PDT) Received: from ulc199.residence.gatech.edu (ulc199.residence.gatech.edu [199.77.162.99]) by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id RAA17655 for ; Thu, 11 Jul 1996 17:57:38 -0700 (PDT) Received: (from ken@localhost) by ulc199.residence.gatech.edu (8.7.5/8.7.3) id UAA08171 Thu, 11 Jul 1996 20:55:58 -0400 (EDT) From: Kenneth Merry Message-Id: <199607120055.UAA08171@ulc199.residence.gatech.edu> Subject: Re: Uptime report To: imp@village.org (Warner Losh) Date: Thu, 11 Jul 1996 20:55:57 -0400 (EDT) Cc: andrew@pubnix.net, stable@FreeBSD.ORG In-Reply-To: <199607112304.RAA21510@rover.village.org> from Warner Losh at "Jul 11, 96 05:04:07 pm" X-Mailer: ELM [version 2.4ME+ PL15 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-stable@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk > : Solution, quit X every so often, all your memory comes back. > > *NO* the memory didn't come back when I did this. The X server was > only about 8M when I exited it. X is not running and I still have 33M > of swap in use on my machine, just like before I exited X. How long has your X server been running? That's the real key, I think. When you run xdm, and log in and out, I don't think the X server process is ever actually killed. It just kinda restarts. I have seen similar symptoms running xdm with AcceleratedX 1.2. The 'fix' is to kill the X server every once in a while, when swap usage gets up there, and the swap space will be returned. (to kill the server, control-alt-bs) xdm will just crank the server right back up. This could be a vm-type problem/feature, in that the memory used by the process is left allocated by the vm system in case it is needed again by the same process. Who knows. > : One could then deduce that the leak is somewhere in X11R6. > > I don't think there is a leak in the X server. However, there may be > sub-optimal memory usage in the X server that causes it to keep > growning and growing due to increaing memory fragmentation. I know a > few people that have purified the X server and found only marginal > leaks (like on the order of 20k over several hours of running > netscape-like programs). Yeah, that kinda seems to point to it being something other than the X server at work. Not sure what that something might be, though. Ken -- Kenneth Merry ken@ulc199.residence.gatech.edu Disclaimer: I don't speak for GTRI, GT, or Elvis.