Re: possible leak in kernel 2.6.10-ac12

[Pedro Venda <pjvenda@arrakis.dhis.org>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Pedro Venda wrote:
| Noel Maddy wrote:
| | On Thu, Feb 17, 2005 at 12:02:33AM +0000, Pedro Venda wrote:
| |
| |
| |>admin proc # cat slabinfo
| |
| | ...
| |
| |>biovec-1           74224  74354     16  226    1 : tunables  120
| 60    0 :
| slabdata    329    329      0
| |>bio                74212  74237     64   61    1 : tunables  120
| 60    0 :
| slabdata   1217   1217      0
| |
| |
| | If you're using md, you need this patch to fix a bio leak:
| |
| | http://linux.bkbits.net:8080/linux-2.6/diffs/drivers/md/md.c@1.234
|
| thanks.
|
| has this patch been included in the recent rc's?
|
| and howcome the other servers work normally [all with md]? at least so
| far the
| memory usages aren't concerning, although the linear increase is
| starting to
| show. perhaps the [different] apps running on them don't expose the leak
| as well
| the one that broke today... is that reasonable?

correction: after looking at /proc/slabinfo, the counters around the bio lines
are 10-100 times bigger than all the other, so their also leaking.

I read some discussion on this list around this issue and people were not
positive the leak came from md...

any comments?

regards.
- --

Pedro João Lopes Venda
email: pjvenda@arrakis.dhis.org
http://arrakis.dhis.org
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.6 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFCFATzeRy7HWZxjWERApHtAKDsX1UVnKP1JICe1/RrzuzF00P84QCfac4+
Mtxq0Y31MFhdsIgBf5S0n/Y=
=u/li
-----END PGP SIGNATURE-----