From: Andrew Morton <akpm@osdl.org>
To: "Sergey S. Kostyliov" <rathamahata@php4.ru>
Cc: linux-kernel@vger.kernel.org, gluk@php4.ru, anton@megashop.ru,
mfedyk@matchmail.com
Subject: Re: 2.6.1 IO lockup on SMP systems
Date: Thu, 26 Feb 2004 04:53:31 -0800 [thread overview]
Message-ID: <20040226045331.060c07d3.akpm@osdl.org> (raw)
In-Reply-To: <200402261519.35506.rathamahata@php4.ru>
"Sergey S. Kostyliov" <rathamahata@php4.ru> wrote:
>
> Yet another lockup has just occurred. I could be wrong but from the
> /proc/meminfo content it doesn't looks like memory leak (neither kernel
> nor userspace), doesn't it?
I think it's a kernel leak.
> Thu Feb 26 05:00:15 MSK 2004
> MemTotal: 2073868 kB
> MemFree: 2528 kB
> Buffers: 2180 kB
> Cached: 34216 kB
> SwapCached: 643808 kB
> Active: 999316 kB
> Inactive: 12088 kB
> HighTotal: 1179648 kB
> HighFree: 576 kB
> LowTotal: 894220 kB
> LowFree: 1952 kB
> SwapTotal: 3583968 kB
> SwapFree: 2559796 kB
> Dirty: 0 kB
> Writeback: 3052 kB
> Mapped: 1001208 kB
> Slab: 23932 kB
> Committed_AS: 1979784 kB
> PageTables: 4840 kB
> VmallocTotal: 114680 kB
> VmallocUsed: 7448 kB
> VmallocChunk: 107232 kB
A gig of mapped memory, most of it in swapcache. That's probably all
highmem. Only a gig of memory on the page LRU. Where is the rest? Lost.
Almost no pagecache at all, slab is small.
> 3) sysrq-T:
> ===========
> http://sysadminday.org.ru/2.6.3-lockup/20040226/sysrq-T
hm, you have 34 instances of crond running. How odd.
> 3) `vmstat 30':
> ===============
> procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
> r b swpd free buff cache si so bi bo in cs us sy id wa
> 0 19 1255096 1952 1996 19920 426 1763 505 1778 1068 172 0 1 0 99
> 0 24 1260156 1944 2028 19816 374 1650 463 1670 1067 165 0 1 0 99
Again, all your memory has vanished.
I'd say that we've leaked everything in lowmem and everyone is stuck trying
to reclaim some lowmem memory. Not sure why the oom-killer didn't do
anything. I haven't tested it in a year - maybe it broke.
So. What are you using which is different from everyone else? DAC960 I
see. What about firewall setups, NIC drivers, RAID/MD/etc? Anything in
there which isn't a mainstream thing?
next prev parent reply other threads:[~2004-02-26 12:53 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2004-01-31 16:40 2.6.1 IO lockup on SMP systems Sergey S. Kostyliov
2004-02-01 0:17 ` Andrew Morton
2004-02-21 16:45 ` Sergey S. Kostyliov
2004-02-21 19:30 ` Andrew Morton
2004-02-22 17:39 ` Alexander Y. Fomichev
2004-02-23 17:27 ` Sergey S. Kostyliov
2004-02-23 21:30 ` Mike Fedyk
2004-02-24 11:56 ` Sergey S. Kostyliov
2004-02-23 22:26 ` Andrew Morton
2004-02-24 7:23 ` Marcelo Tosatti
2004-02-24 6:53 ` Andrew Morton
2004-02-24 11:54 ` Sergey S. Kostyliov
2004-02-26 12:19 ` Sergey S. Kostyliov
2004-02-26 12:53 ` Andrew Morton [this message]
2004-02-26 13:11 ` Andrew Morton
2004-02-26 14:37 ` Dave Jones
2004-02-26 15:37 ` Arjan van de Ven
2004-02-26 14:30 ` Sergey S. Kostyliov
2004-02-26 20:03 ` Andrew Morton
2004-02-28 14:56 ` Sergey S. Kostyliov
2004-04-08 9:08 ` 2.6.X kernel memory leak? (was: Re: 2.6.1 IO lockup on SMP systems) Sergey S. Kostyliov
2004-04-09 7:17 ` 2.6.X kernel memory leak? Sergey S. Kostyliov
2004-04-09 9:09 ` Andrew Morton
2004-04-09 12:15 ` Sergey S. Kostyliov
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=20040226045331.060c07d3.akpm@osdl.org \
--to=akpm@osdl.org \
--cc=anton@megashop.ru \
--cc=gluk@php4.ru \
--cc=linux-kernel@vger.kernel.org \
--cc=mfedyk@matchmail.com \
--cc=rathamahata@php4.ru \
/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