From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jeff Mahoney Subject: Re: ReiserFS causes CPU high usage?! Date: Fri, 28 Feb 2003 11:43:42 -0500 Message-ID: <3E5F91BE.5050707@suse.com> References: <200302280011.14488.liviu.voicu@huji.ac.il> <200302281439.24760.liviu.voicu@huji.ac.il> <20030228133836.D24291@humilis> <200302281448.20572.liviu.voicu@huji.ac.il> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Return-path: list-help: list-unsubscribe: list-post: Errors-To: flx@namesys.com In-Reply-To: <200302281448.20572.liviu.voicu@huji.ac.il> List-Id: Content-Type: text/plain; charset="us-ascii"; format="flowed" To: Voicu Liviu Cc: ookhoi@humilis.net, reiserfs-list@namesys.com Voicu Liviu wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > OK, so in this mashine I compile kde3.1 and this is the output of the "top" > command ( it took me 5 seconds to write top in the command line and another > 10 antill I saw the output...really slow ) > > 35 processes: 31 sleeping, 4 running, 0 zombie, 0 stopped > CPU states: 99.4% user, 0.3% system, 0.0% nice, 0.0% iowait, 0.1% idle If ReiserFS were causing your CPU utilization problems, it would look more like this: CPU0 states: 0.5% user, 28.1% system, 0.0% nice, 70.3% idle CPU1 states: 0.3% user, 28.1% system, 0.0% nice, 71.0% idle This is the result when I start a stress.sh -n 50 run, which starts 50 processes all copying a directory to a ReiserFS filesystem. [dual p3 733] The numbers vary, but the system number is usually higher than just a few percent. Your system time is only 0.3%, but your user time is 99.4%. It's your userland that's taking all your CPU time. There's a large number of -10 nice'd processes in there: The result is that those processes get scheduler priority, and things that you're looking for interactive response from won't get it. I noticed that someone further down the list mentioned that your slowdown may be sshd not getting enough CPU time. This seems much more likely. To answer your other question about ReiserFS on a p2 300: I have a closet server running with a few hundred GB of ReiserFS filesystems on it, acting as an NFS/Samba server, handling internal web traffic, etc.. It's got cycles to spare, and it's only a Celeron 300A. -Jeff > Mem: 255948k av, 250540k used, 5408k free, 0k shrd, 28344k buff > 59512k active, 119096k inactive > Swap: 514072k av, 144k used, 513928k free 122280k > cached > > PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND > 889 portage 15 -10 21508 21M 4172 R < 98.9 8.4 0:16 cc1plus > 890 portage 5 -10 2368 2368 752 S < 0.9 0.9 0:00 as > 845 liviu 15 0 1100 1100 792 R 0.1 0.4 0:00 top > 1 root 15 0 484 484 428 S 0.0 0.1 0:05 init > 2 root 15 0 0 0 0 SW 0.0 0.0 0:00 keventd > 3 root 34 19 0 0 0 SWN 0.0 0.0 0:00 ksoftirqd_CPU0 > 4 root 15 0 0 0 0 SW 0.0 0.0 0:00 kswapd > 5 root 15 0 0 0 0 SW 0.0 0.0 0:00 bdflush > 6 root 15 0 0 0 0 RW 0.0 0.0 0:00 kupdated > 9 root 15 0 0 0 0 SW 0.0 0.0 0:00 kreiserfsd > 33 root 15 0 888 888 592 S 0.0 0.3 0:00 devfsd > 3612 root 15 0 564 564 488 S 0.0 0.2 0:00 metalog > 3614 root 15 0 460 460 416 S 0.0 0.1 0:00 metalog > 3729 root 15 0 1252 1252 1128 S 0.0 0.4 0:00 sshd > 3740 root 15 0 412 412 356 S 0.0 0.1 0:00 agetty > 3741 root 15 0 412 412 356 S 0.0 0.1 0:00 agetty > 3742 root 15 0 412 412 356 S 0.0 0.1 0:00 agetty > 3743 root 15 0 412 412 356 S 0.0 0.1 0:00 agetty > 3744 root 15 0 412 412 356 S 0.0 0.1 0:00 agetty > 3745 root 15 0 412 412 356 S 0.0 0.1 0:00 agetty > 12092 root 15 0 1220 1220 836 S 0.0 0.4 0:00 screen > 12093 root 15 0 1316 1316 1044 S 0.0 0.5 0:00 sh > 12099 root 10 -5 5492 5492 1896 S < 0.0 2.1 0:07 python2.2 > 31549 portage 10 -10 376 376 312 S < 0.0 0.1 0:00 sandbox > 31550 portage 10 -10 1172 1172 900 S < 0.0 0.4 0:00 ebuild.sh > 31557 portage 5 -10 472 472 396 S < 0.0 0.1 0:00 tee > 31577 liviu 15 0 1316 1316 1060 S 0.0 0.5 0:00 bash > 300 portage 10 -10 796 796 556 S < 0.0 0.3 0:00 make > 699 portage 10 -10 832 832 576 S < 0.0 0.3 0:00 make > 700 portage 10 -10 1052 1052 836 S < 0.0 0.4 0:00 sh > 701 portage 6 -10 1552 1552 576 S < 0.0 0.6 0:00 make > 886 portage 6 -10 640 640 440 S < 0.0 0.2 0:00 g++ > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v1.2.1 (GNU/Linux) > > iD8DBQE+X1qUczD4Ii52ssMRAso7AJ0bAlC3YbXjax28nVe9NbI6/+cJ/ACfX9Lx > O+fZu4gbCPjqSDraeWrYbbk= > =uWh1 > -----END PGP SIGNATURE----- -- jeffm@suse.com jeffm@csh.rit.edu