On Wed, 10 Dec 2003, Konstantin Nikonenko wrote:

> Hello oops-users,
>
> Tuesday, December 9, 2003, 7:07:03 PM, you wrote:
> LEK> We have a problem with oops (both 1.5.23 and the cvs version), it died
> LEK> a couple hours after it has been started.
> LEK> Here is the output of oops.log:
>
> LEK> [0x19465]run_client(): No mem for header!
> LEK> [0x6b1ac]run_client(): No mem for header!
> LEK> [0x9a268]fill_mem_obj(): select: timed out.
> LEK> [0xbcef3]attach_data(): No mem in attach data.
> LEK> [0x5795e]run_client(): No mem for header!
>
> What gdb is answer on where in core? How many memory oops use? mb
No, it just dies without core dumped.
Output from top several minutes before it dies:

top - 18:31:05 up 18:36,  2 users,  load average: 0.57, 0.38, 0.36
Tasks: 699 total,   1 running, 698 sleeping,   0 stopped,   0 zombie
Cpu(s):  5.6% us,  2.1% sy,  0.0% ni, 89.4% id,  0.6% wa,  0.1% hi,  2.2% si
Mem:   3111636k total,  3025188k used,    86448k free,   239164k buffers
Swap:  2097136k total,        0k used,  2097136k free,  1385288k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
70858 oopscach  17   0 2417m 1.1g 3604 S 21.6 37.8   1:07.28 oops
70869 oopscach  15   0 2417m 1.1g 3604 S 21.6 37.8   1:07.28 oops
[..more output omitted..]

We've set the mem_max and lo_mark to 1024m,
but it continue growing after it reached the limit.

> needs set limits or more RAM|swap?
ulimit:
core file size        (blocks, -c) 0
data seg size         (kbytes, -d) unlimited
file size             (blocks, -f) unlimited
max locked memory     (kbytes, -l) unlimited
max memory size       (kbytes, -m) unlimited
open files                    (-n) 8192
pipe size          (512 bytes, -p) 8
stack size            (kbytes, -s) 8192
cpu time             (seconds, -t) unlimited
max user processes            (-u) 24575
virtual memory        (kbytes, -v) unlimited

>
> LEK> There's no other messages abnormal, but sometimes with below messages
> LEK> from kernel (not always):
>
> LEK> Dec  8 13:54:05 tproxy1 kernel: oops: page allocation failure. order:3, 
> mode:0x20
> LEK> Dec  8 13:54:05 tproxy1 kernel: Call Trace:
> LEK> Dec  8 13:54:05 tproxy1 kernel:  [<c013ac0d>] __alloc_pages+0x2de/0x31c
> LEK> Dec  8 13:54:05 tproxy1 kernel:  [<c013ac70>] __get_free_pages+0x25/0x3f
> LEK> Dec  8 13:54:05 tproxy1 kernel:  [<c013e0cd>] cache_grow+0xed/0x345
>
> LEK> The swap remain unused at that time.
>
> LEK> Tested on both kernel 2.4.23 and 2.6.0-test9-mm5,
> LEK> glibc-2.2.5, gcc-3.3.2 (gcc-2.95.3 also), Debian GNU/Linux-3.0.
> what threads type are u use?
$ gcc -v
Reading specs from /usr/local/lib/gcc-lib/i686-pc-linux-gnu/3.3.2/specs
Configured with: ./configure --prefix=/usr/local
Thread model: posix
gcc version 3.3.2
$
>
> LEK> The request rate varies from 100 req/s to 550 req/s.
> what hardware u use? IMHO for this requests rate Linux not a best
> choise :( MB memory manager really can't allocate memory on this
> speed...
hmm..
We're running on Dual Xeon 2.8G with HT enabled, 3GB of RAM and the
chipset is intel E7501.
The kernel is configured with HIGHMEM support(4GB).
>
> LEK> Any ideas what's wrong?
>
> --
> Best regards,
>  Konstantin Nikonenko        http://www.kot.dp.ua/
>
> =====================================================================
> If you would like to unsubscribe from this list send message to
> [EMAIL PROTECTED] with "unsubscribe oops" in message body.
> Archive is accessible on http://lists.paco.net/oops-rus/
>
=====================================================================
If you would like to unsubscribe from this list send message to
[EMAIL PROTECTED] with "unsubscribe oops" in message body.
Archive is accessible on http://lists.paco.net/oops-rus/

Дати відповідь електронним листом