* [PROBLEM] Badness in out_of_memory
@ 2005-01-06 17:03 Richard Ems
2005-01-06 17:30 ` Andrea Arcangeli
0 siblings, 1 reply; 5+ messages in thread
From: Richard Ems @ 2005-01-06 17:03 UTC (permalink / raw)
To: linux-kernel, Hubert Mantel, andrea
Hi list, hi Mr. Mantel,
the following "badness" happened on a SuSE 9.2 with all actual updates
and SuSE's kernel 2.6.8-24.10-smp.
The system is a dual AMD Athlon MP 2200+ with 1GB memory and 1GB swap.
# uname -a
Linux urutu 2.6.8-24.10-smp #1 SMP Wed Dec 22 11:54:27 UTC 2004 i686
athlon i386 GNU/Linux
# free
total used free shared buffers cached
Mem: 1034404 683964 350440 0 30584 412916
-/+ buffers/cache: 240464 793940
Swap: 1052248 121900 930348
In /var/log/messages I found 161 times the following "badness" lines.
The system load rised up to over 100, 159 from the "badnesses" where
logged on the same second in /var/log/messages.
# grep -i badness /var/log/messages|wc -l
161
# grep -i badness /var/log/messages | head -5
Jan 6 14:42:24 urutu kernel: Badness in out_of_memory at mm/oom_kill.c:252
Jan 6 14:42:30 urutu kernel: Badness in out_of_memory at mm/oom_kill.c:252
Jan 6 14:42:31 urutu kernel: Badness in out_of_memory at mm/oom_kill.c:252
Jan 6 14:42:31 urutu kernel: Badness in out_of_memory at mm/oom_kill.c:252
Jan 6 14:42:31 urutu kernel: Badness in out_of_memory at mm/oom_kill.c:252
# grep -i badness /var/log/messages | tail -3
Jan 6 14:42:31 urutu kernel: Badness in out_of_memory at mm/oom_kill.c:252
Jan 6 14:42:31 urutu kernel: Badness in out_of_memory at mm/oom_kill.c:252
Jan 6 14:42:31 urutu kernel: Badness in out_of_memory at mm/oom_kill.c:252
Jan 6 14:42:24 urutu kernel: Badness in out_of_memory at mm/oom_kill.c:252
Jan 6 14:42:27 urutu kernel: [<c01464a9>] out_of_memory+0x39/0x130
Jan 6 14:42:27 urutu kernel: [<c014f686>] try_to_free_pages+0x196/0x1a0
Jan 6 14:42:27 urutu kernel: [<c0147800>] __alloc_pages+0x290/0x400
Jan 6 14:42:27 urutu kernel: [<f8856d7c>]
dm_table_unplug_all+0x2c/0x50 [dm_mod]
Jan 6 14:42:27 urutu kernel: [<c014a82c>]
do_page_cache_readahead+0x13c/0x180
Jan 6 14:42:27 urutu kernel: [<c0144139>] filemap_nopage+0x299/0x370
Jan 6 14:42:27 urutu kernel: [<c0153022>] do_no_page+0xd2/0x2e0
Jan 6 14:42:27 urutu kernel: [<c015348d>] handle_mm_fault+0x15d/0x1b0
Jan 6 14:42:27 urutu kernel: [<c011c403>] do_page_fault+0x1d3/0x5ee
Jan 6 14:42:27 urutu kernel: [<c01754d0>] __pollwait+0x0/0xd0
Jan 6 14:42:27 urutu kernel: [<c010f1f5>] convert_fxsr_from_user+0x15/0xf0
Jan 6 14:42:27 urutu kernel: [<c0175bfd>] sys_select+0x23d/0x4c0
Jan 6 14:42:28 urutu kernel: [<c01f6de2>] copy_to_user+0x32/0x50
Jan 6 14:42:28 urutu kernel: [<c011c230>] do_page_fault+0x0/0x5ee
Jan 6 14:42:28 urutu kernel: [<c01081bd>] error_code+0x2d/0x40
Jan 6 14:42:30 urutu kernel: x1d3/0x5ee
Jan 6 14:42:30 urutu kernel: [<c029bb78>] start_request+0x278/0x2c0
Jan 6 14:42:30 urutu kernel: [<c010f1f5>] convert_fxsr_from_user+0x15/0xf0
Jan 6 14:42:30 urutu kernel: [<c010f537>] restore_i387+0x67/0x70
Jan 6 14:42:30 urutu kernel: [<c0106404>] restore_sigcontext+0x114/0x130
Jan 6 14:42:30 urutu kernel: [<c01064e8>] sys_sigreturn+0xc8/0xe0
Jan 6 14:42:30 urutu kernel: [<c011c230>] do_page_fault+0x0/0x5ee
Jan 6 14:42:30 urutu kernel: [<c01081bd>] error_code+0x2d/0x40
The process killed were:
# grep Killed /var/log/messages
Jan 6 14:42:31 urutu kernel: Out of Memory: Killed process 26123 (java_vm).
Jan 6 14:42:31 urutu kernel: Out of Memory: Killed process 26126 (java_vm).
Jan 6 14:42:31 urutu kernel: Out of Memory: Killed process 26120 (java_vm).
Jan 6 14:42:31 urutu kernel: Out of Memory: Killed process 9730
(mozilla-bin).
Jan 6 14:42:31 urutu kernel: Out of Memory: Killed process 9732
(mozilla-bin).
Jan 6 14:42:31 urutu kernel: Out of Memory: Killed process 9739
(mozilla-bin).
Any idea what's happening here? Do you need more info?
Thanks, Richard
--
Richard Ems
Tel: +49 40 65803 312
Fax: +49 40 65803 392
Richard.Ems@mtg-marinetechnik.de
MTG Marinetechnik GmbH - Wandsbeker Königstr. 62 - D 22041 Hamburg
GF Dipl.-Ing. Ullrich Keil
Handelsregister: Abt. B Nr. 11 500 - Amtsgericht Hamburg Abt. 66
USt.-IdNr.: DE 1186 70571
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PROBLEM] Badness in out_of_memory
2005-01-06 17:03 [PROBLEM] Badness in out_of_memory Richard Ems
@ 2005-01-06 17:30 ` Andrea Arcangeli
2005-01-06 17:51 ` [PROBLEM] Badness in out_of_memory (Plain) Richard Ems
2005-01-07 15:18 ` Richard Ems
0 siblings, 2 replies; 5+ messages in thread
From: Andrea Arcangeli @ 2005-01-06 17:30 UTC (permalink / raw)
To: Richard Ems; +Cc: linux-kernel, Hubert Mantel
On Thu, Jan 06, 2005 at 06:03:35PM +0100, Richard Ems wrote:
> Hi list, hi Mr. Mantel,
>
> the following "badness" happened on a SuSE 9.2 with all actual updates
> and SuSE's kernel 2.6.8-24.10-smp.
> The system is a dual AMD Athlon MP 2200+ with 1GB memory and 1GB swap.
This is a warning only (2.6.9 had the swap token breakage that triggered
suprious oom kills, so the warning was meant to get more info), can you
try with the kernel of the day?
http://ftp.suse.com/pub/projects/kernel/kotd/9.2-i386/SL92_BRANCH
It has my latest oom fixes that I recently posted to l-k and it should
be very reliable for the first time in oom-killer terms.
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PROBLEM] Badness in out_of_memory (Plain)
2005-01-06 17:30 ` Andrea Arcangeli
@ 2005-01-06 17:51 ` Richard Ems
2005-01-06 23:41 ` Andrea Arcangeli
2005-01-07 15:18 ` Richard Ems
1 sibling, 1 reply; 5+ messages in thread
From: Richard Ems @ 2005-01-06 17:51 UTC (permalink / raw)
To: Andrea Arcangeli; +Cc: linux-kernel, Hubert Mantel
Andrea Arcangeli wrote:
> On Thu, Jan 06, 2005 at 06:03:35PM +0100, Richard Ems wrote:
>
>>Hi list, hi Mr. Mantel,
>>
>>the following "badness" happened on a SuSE 9.2 with all actual updates
>>and SuSE's kernel 2.6.8-24.10-smp.
>>The system is a dual AMD Athlon MP 2200+ with 1GB memory and 1GB swap.
>
>
> This is a warning only (2.6.9 had the swap token breakage that triggered
> suprious oom kills, so the warning was meant to get more info), can you
> try with the kernel of the day?
>
> http://ftp.suse.com/pub/projects/kernel/kotd/9.2-i386/SL92_BRANCH
>
> It has my latest oom fixes that I recently posted to l-k and it should
> be very reliable for the first time in oom-killer terms.
>
>
I will try it tomorrow evening, it's a production system, so it's not so
easy to "try".
How can I trigger the oom kills? I'm not sure if I can easily reproduce
this! Do you know a memory eater program other than starting lots of
mozilla's, OpenOffices's, etc.?
Thanks, Richard
--
Richard Ems
Tel: +49 40 65803 312
Fax: +49 40 65803 392
Richard.Ems@mtg-marinetechnik.de
MTG Marinetechnik GmbH - Wandsbeker Königstr. 62 - D 22041 Hamburg
GF Dipl.-Ing. Ullrich Keil
Handelsregister: Abt. B Nr. 11 500 - Amtsgericht Hamburg Abt. 66
USt.-IdNr.: DE 1186 70571
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PROBLEM] Badness in out_of_memory (Plain)
2005-01-06 17:51 ` [PROBLEM] Badness in out_of_memory (Plain) Richard Ems
@ 2005-01-06 23:41 ` Andrea Arcangeli
0 siblings, 0 replies; 5+ messages in thread
From: Andrea Arcangeli @ 2005-01-06 23:41 UTC (permalink / raw)
To: Richard Ems; +Cc: linux-kernel, Hubert Mantel
On Thu, Jan 06, 2005 at 06:51:18PM +0100, Richard Ems wrote:
> this! Do you know a memory eater program other than starting lots of
> mozilla's, OpenOffices's, etc.?
You can run a few of this.
#include <stdio.h>
#include <time.h>
main()
{
char *p[160];
int i, j;
int count;
for (j=0; j<160; j++)
{
p[j] = (char *) malloc(10000000);
}
for (count=0;count<2000;count++)
{
for (j=0; j<160; j++)
{
for (i=0; i<10000000; i+= 4096)
p[j][i] = 0;
}
pause();
}
}
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PROBLEM] Badness in out_of_memory (Plain)
2005-01-06 17:30 ` Andrea Arcangeli
2005-01-06 17:51 ` [PROBLEM] Badness in out_of_memory (Plain) Richard Ems
@ 2005-01-07 15:18 ` Richard Ems
1 sibling, 0 replies; 5+ messages in thread
From: Richard Ems @ 2005-01-07 15:18 UTC (permalink / raw)
To: Andrea Arcangeli; +Cc: linux-kernel, Hubert Mantel
Andrea Arcangeli wrote:
> This is a warning only (2.6.9 had the swap token breakage that triggered
> suprious oom kills, so the warning was meant to get more info), can you
> try with the kernel of the day?
>
> http://ftp.suse.com/pub/projects/kernel/kotd/9.2-i386/SL92_BRANCH
>
> It has my latest oom fixes that I recently posted to l-k and it should
> be very reliable for the first time in oom-killer terms.
I tried it and it is working ok, when memory and swap become full
processes are being killed very fast and the system becomes responsive
again.
Do you want me to do some specific test and send you some results?
Thanks, Richard
--
Richard Ems
Tel: +49 40 65803 312
Fax: +49 40 65803 392
Richard.Ems@mtg-marinetechnik.de
MTG Marinetechnik GmbH - Wandsbeker Königstr. 62 - D 22041 Hamburg
GF Dipl.-Ing. Ullrich Keil
Handelsregister: Abt. B Nr. 11 500 - Amtsgericht Hamburg Abt. 66
USt.-IdNr.: DE 1186 70571
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2005-01-07 15:18 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2005-01-06 17:03 [PROBLEM] Badness in out_of_memory Richard Ems
2005-01-06 17:30 ` Andrea Arcangeli
2005-01-06 17:51 ` [PROBLEM] Badness in out_of_memory (Plain) Richard Ems
2005-01-06 23:41 ` Andrea Arcangeli
2005-01-07 15:18 ` Richard Ems
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®