Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Wed, 7 Nov 2001 15:04:00 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Wed, 7 Nov 2001 15:03:50 -0500 Received: from fe5.southeast.rr.com ([24.93.67.52]:31241 "EHLO mail5.nc.rr.com") by vger.kernel.org with ESMTP id ; Wed, 7 Nov 2001 15:03:40 -0500 Subject: Re: Linux-2.4.[10-13]-acX slows down to crawl... From: "C. Linus Hicks" To: Andreas.Franck@akustik.rwth-aachen.de Cc: linux-kernel@vger.kernel.org In-Reply-To: <3BE95E9C.D62D372C@akustik.rwth-aachen.de> In-Reply-To: <3BE95E9C.D62D372C@akustik.rwth-aachen.de> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/0.12 (Preview Release) Date: 07 Nov 2001 12:29:33 -0500 Message-Id: <1005154173.2537.37.camel@lh2> Mime-Version: 1.0 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On 07 Nov 2001 17:17:32 +0100, Andreas Franck wrote: > Hello folks, > > since quite a time (it happens at least since 2.4.10-ac, and also > happens with the newest -ac kernel 2.4.13-ac8), Linux gets slower > and slower, to a state where it is basically unusable after some days. > I suspect it is a VM issue, because it starts swapping without any > need, not freeing buffers? The buffer sizes are about the size of RAM > when it is worst. For anything done, the load goes up extremely, network > access basically halts the machine then, leading to a load of about 20 > :-( I did not look closely at what was happening, but I experienced a similar slowdown on 2.4.9-ac11. It took a while in my case, (several weeks) as I have 1gb memory and don't have any massive files on my system. I have dual PIII 1GHz CPUs and was downloading a 300mb file when I started xpdf on a document that has lots of complex drawaings. The download rate was 200-250kbps prior to starting xpdf and dropped to 20-40. If I stopped the download and restarted, it ran at 200-250 until I ran xpdf again (that it was running did not seem important, but when I displayed a new page and it ate a CPU). I thought it might be something with xpdf because I was able to reproduce an xpdf slowdown by running three copies at the same time. However, now I am running 2.4.12-ac6 booted last night and I have seen the slowdown in xpdf running just one. Last night after I re-booted, it was running fine. The strange thing is that when this happens, the system acts like it has no spare CPU because even switching to another window takes a long time for the focus to change. top reports: 12:21pm up 11:36, 16 users, load average: 0.30, 0.13, 0.09 140 processes: 138 sleeping, 2 running, 0 zombie, 0 stopped CPU0 states: 0.5% user, 0.4% system, 0.0% nice, 98.1% idle CPU1 states: 99.3% user, 0.2% system, 0.0% nice, 0.0% idle Mem: 1028936K av, 817768K used, 211168K free, 4152K shrd, 282336K buff Swap: 2000880K av, 0K used, 2000880K free 189408K cached PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND 1117 root 18 0 75864 34M 5972 R 99.9 3.3 342:36 X 5491 linush 14 0 1016 1016 768 R 1.1 0.0 0:09 top 4859 linush 9 0 3544 3544 1616 S 0.5 0.3 0:04 xpdf 1 root 8 0 528 528 460 S 0.0 0.0 0:01 init 2 root 9 0 0 0 0 SW 0.0 0.0 0:00 keventd Let me know if I can be of any assistance. Linus - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/