From: kladit@t-online.de (Klaus Dittrich)
To: linux mailing-list <linux-kernel@vger.kernel.org>
Cc: kladit@t-online.de
Subject: Re: dentry cache leak? Re: rsync out of memory 2.6.8-rc2
Date: Mon, 26 Jul 2004 17:06:15 +0200 [thread overview]
Message-ID: <20040726150615.GA1119@xeon2.local.here> (raw)
>Can you narrow the onset of the problem down to any particular kernel
>snapshot?
Did it and here is the answer.
kernel-2.6.7 and bk's up to 2.6.7-bk7 survived a du -s,
kernels starting with 2.6.7-bk8 did not.
Compiler gcc-3.4.1
System SMP 2 Xeon-CPU's
Hardware Tyan 2665 Mobo
/disc1 is a software raid-1, mirror of two 60GB IDE disks.
To be sure to have clean conditions, I recompiled 2.6.7 and
some subsequent -bk kernels.
The configuration of each kernel is of a copy of .config from
2.6.7 and innovations during make oldconfig were answered using
the defaults.
Some observations while du -s is running:
With the 'good' kernels there is a freeze of about a second
when memory usage (xosview) has grown up to about 1.4 GByte.
Not using irqbalance results in very high system load of CPU0.
Gathered data follows below. Please cc me.
--
Klaus
linux-2.6.7 (good)
===========
cat /proc/sys/fs/dentry-state
7015 4538 45 0 0 0
du -s /disc1/
42119916 /disc1/
cat /proc/sys/fs/dentry-state
719959 690331 45 0 0 0
cat /proc/vmstat
nr_dirty 6
nr_writeback 0
nr_unstable 0
nr_page_table_pages 417
nr_mapped 36574
nr_slab 219486
pgpgin 822287
pgpgout 125952
pswpin 0
pswpout 0
pgalloc_high 218904
pgalloc_normal 590088
pgalloc_dma 3534
pgfree 967065
pgactivate 182625
pgdeactivate 156250
pgfault 216065
pgmajfault 2694
pgrefill_high 0
pgrefill_normal 154846
pgrefill_dma 1404
pgsteal_high 0
pgsteal_normal 81532
pgsteal_dma 483
pgscan_kswapd_high 0
pgscan_kswapd_normal 130065
pgscan_kswapd_dma 721
pgscan_direct_high 0
pgscan_direct_normal 179566
pgscan_direct_dma 2425
pginodesteal 12160
slabs_scanned 2817088
kswapd_steal 51850
kswapd_inodesteal 1888
pageoutrun 1083
allocstall 2857
pgrotated 8111
linux-2.6.7-bk6 (good)
===============
du -s /disc1/
42120684 /disc1/
cat /proc/sys/fs/dentry-state
720795 691157 45 0 0 0
cat /proc/vmstat
nr_dirty 11
nr_writeback 0
nr_unstable 0
nr_page_table_pages 459
nr_mapped 42928
nr_slab 219592
pgpgin 848899
pgpgout 135604
pswpin 0
pswpout 0
pgalloc_high 256670
pgalloc_normal 619555
pgalloc_dma 3210
pgfree 1023837
pgactivate 185293
pgdeactivate 157150
pgfault 264830
pgmajfault 3024
pgrefill_high 0
pgrefill_normal 156851
pgrefill_dma 299
pgsteal_high 0
pgsteal_normal 82775
pgsteal_dma 397
pgscan_kswapd_high 0
pgscan_kswapd_normal 126659
pgscan_kswapd_dma 457
pgscan_direct_high 0
pgscan_direct_normal 147678
pgscan_direct_dma 365
pginodesteal 6491
slabs_scanned 2800364
kswapd_steal 55007
kswapd_inodesteal 8269
pageoutrun 1147
allocstall 2225
pgrotated 8831
linux-2.6.7-bk7 (good)
===============
cat /proc/sys/fs/dentry-state
6990 4522 45 0 0 0
du -s /disc1
..
some snapshoots of /proc/sys/fs/dentry-state
761894 723814 45 0 0 0
760283 722353 45 0 0 0
759821 721937 45 0 0 0
759872 722127 45 0 0 0
759742 722079 45 0 0 0
759666 722090 45 0 0 0
757161 719829 45 0 0 0
755236 718111 45 0 0 0
755639 718584 45 0 0 0
714917 683972 45 0 0 0
716095 685185 45 0 0 0
715329 684496 45 0 0 0
..
42121188 /disc1
cat /proc/vmstat
nr_dirty 7
nr_writeback 0
nr_unstable 0
nr_page_table_pages 469
nr_mapped 43769
nr_slab 219515
pgpgin 854375
pgpgout 136916
pswpin 0
pswpout 0
pgalloc_high 265027
pgalloc_normal 621919
pgalloc_dma 3790
pgfree 1031340
pgactivate 196742
pgdeactivate 169483
pgfault 288603
pgmajfault 3068
pgrefill_high 0
pgrefill_normal 169001
pgrefill_dma 482
pgsteal_high 0
pgsteal_normal 83976
pgsteal_dma 490
pgscan_kswapd_high 0
pgscan_kswapd_normal 134412
pgscan_kswapd_dma 738
pgscan_direct_high 0
pgscan_direct_normal 175212
pgscan_direct_dma 1310
pginodesteal 3476
slabs_scanned 2826482
kswapd_steal 55611
kswapd_inodesteal 6862
pageoutrun 1214
allocstall 2734
pgrotated 8730
linux-2.6.7-bk8 (bad)
===============
cat /proc/sys/fs/dentry-state
7357 4821 45 0 0 0
du -s /disc1/
After some processes wer killed ..
cat /proc/sys/fs/dentry-state
1090719 1035089 45 0 0 0
cat /proc/vmstat
nr_dirty 26
nr_writeback 0
nr_unstable 0
nr_page_table_pages 365
nr_mapped 29495
nr_slab 219834
pgpgin 594519
pgpgout 80416
pswpin 0
pswpout 0
pgalloc_high 199307
pgalloc_normal 489087
pgalloc_dma 3711
pgfree 868142
pgactivate 94618
pgdeactivate 65790
pgfault 227492
pgmajfault 3065
pgrefill_high 0
pgrefill_normal 65694
pgrefill_dma 96
pgsteal_high 0
pgsteal_normal 46724
pgsteal_dma 337
pgscan_kswapd_high 0
pgscan_kswapd_normal 49351
pgscan_kswapd_dma 331
pgscan_direct_high 0
pgscan_direct_normal 46908
pgscan_direct_dma 41
pginodesteal 2743
slabs_scanned 1186575
kswapd_steal 39703
kswapd_inodesteal 6800
pageoutrun 323
allocstall 1088
pgrotated 1382
next reply other threads:[~2004-07-26 15:34 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2004-07-26 15:06 Klaus Dittrich [this message]
2004-07-26 19:37 ` dentry cache leak? Re: rsync out of memory 2.6.8-rc2 Andrew Morton
2004-07-26 20:02 ` Klaus Dittrich
2004-07-26 20:38 ` Andrew Morton
2004-07-26 21:39 ` Klaus Dittrich
2004-07-26 22:14 ` Jan-Frode Myklebust
2004-07-27 20:43 ` Klaus Dittrich
2004-07-27 21:05 ` Klaus Dittrich
2004-07-28 1:21 ` Nick Piggin
2004-07-28 15:06 ` Klaus Dittrich
2004-07-29 3:31 ` Nick Piggin
2004-07-29 21:07 ` Andrew Morton
2004-07-30 16:30 ` Marcelo Tosatti
2004-07-30 19:47 ` Andrew Morton
2004-07-31 2:01 ` Marcelo Tosatti
2004-07-31 17:23 ` Linus Torvalds
2004-07-31 21:39 ` Andrew Morton
2004-08-03 6:55 ` Nick Piggin
-- strict thread matches above, loose matches on Subject: below --
2004-08-01 9:47 Klaus Dittrich
2004-07-19 9:19 Klaus Dittrich
2004-07-19 11:20 ` dentry cache leak? " bert hubert
[not found] ` <20040719113228.GA15295@outpost.ds9a.nl>
2004-07-19 12:56 ` Klaus Dittrich
2004-07-26 1:06 ` Andrew Morton
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20040726150615.GA1119@xeon2.local.here \
--to=kladit@t-online.de \
--cc=linux-kernel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox