AW: [Linux-ha-dev] Hb-2.0.9/SuSE: memory leak in crmd/lrmd undersolaris 10/i386

[Otte, Joerg]

> are these the only two processes growing?
There are other (smaller) increments as well.
Here is the complete list:

RSS  VSZ    RSS  VSZ
1708 4496   1708 4496  /usr/sfw/lib/python2.3/heartbeat/ha_logd
1316 4396   1316 4396  /usr/sfw/lib/python2.3/heartbeat/ha_logd
9204 12788  9204 12788 /usr/sfw/lib/python2.3/heartbeat/heartbeat
1292 5720   1292 5720  /usr/sfw/lib/python2.3/heartbeat/heartbeat
1500 5776   1500 5776  /usr/sfw/lib/python2.3/heartbeat/heartbeat
1548 5776   1548 5776  /usr/sfw/lib/python2.3/heartbeat/heartbeat
1500 5776   1500 5776  /usr/sfw/lib/python2.3/heartbeat/heartbeat
1524 5776   1524 5776  /usr/sfw/lib/python2.3/heartbeat/heartbeat
2468 4632   2500 4664  /usr/sfw/lib/python2.3/heartbeat/ccm
5484 8524   5496 8524  /usr/sfw/lib/python2.3/heartbeat/cib
5840 7860   6496 8516  /usr/sfw/lib/python2.3/heartbeat/lrmd
2788 4652   2884 4748  /usr/sfw/lib/python2.3/heartbeat/stonithd
2500 4576   2500 4576  /usr/sfw/lib/python2.3/heartbeat/attrd
5716 7580   6488 8356  /usr/sfw/lib/python2.3/heartbeat/crmd
3316 6704   3316 6704  /usr/sfw/lib/python2.3/heartbeat/mgmtd
4476 6500   4480 6500  /usr/sfw/lib/python2.3/heartbeat/tengine
6396 10372  6396 10372 /usr/sfw/lib/python2.3/heartbeat/pengine

> over what interval is the growth?
12 hours

> are there any tools on solaris to show where the leaks are coming
> from? (on linux we have valgrind, support for which is easily enabled
> in heartbeat)
I don't know of any tool which comes for free.


-----Ursprüngliche Nachricht-----
Von: linux-ha-dev-bounces at lists.linux-ha.org [mailto:linux-ha-dev-bounces at lists.linux-ha.org] Im Auftrag von ext Andrew Beekhof
Gesendet: Freitag, 22. Juni 2007 09:42
An: High-Availability Linux Development List
Betreff: Re: [Linux-ha-dev] Hb-2.0.9/SuSE: memory leak in crmd/lrmd undersolaris 10/i386

On 6/22/07, Otte, Joerg <joerg.otte at nsn.com> wrote:
>
>
>
> During a long-time run I am giving a cibadmin command to change
> the preferred location of a resource group once a minute (cibadmin -M -X
> '...').
>
> I trace the used memory with ps and see the memory consumed by
> crmd,lrmd growing permanently:

are these the only two processes growing?
over what interval is the growth?

are there any tools on solaris to show where the leaks are coming
from? (on linux we have valgrind, support for which is easily enabled
in heartbeat)

>
> lrmd:
> RSS  VSZ
> 5840 7860 /usr/sfw/lib/python2.3/heartbeat/lrmd
> 5856 7876 /usr/sfw/lib/python2.3/heartbeat/lrmd
> 5864 7884 /usr/sfw/lib/python2.3/heartbeat/lrmd
> ..
> ..
> 6472 8492 /usr/sfw/lib/python2.3/heartbeat/lrmd
> 6488 8508 /usr/sfw/lib/python2.3/heartbeat/lrmd
> 6496 8516 /usr/sfw/lib/python2.3/heartbeat/lrmd
>
> crmd:
> RSS  VSZ
> 5716 7580 /usr/sfw/lib/python2.3/heartbeat/crmd
> 5724 7588 /usr/sfw/lib/python2.3/heartbeat/crmd
> 5740 7604 /usr/sfw/lib/python2.3/heartbeat/crmd
> ..
> ..
> 6396 8260 /usr/sfw/lib/python2.3/heartbeat/crmd
> 6440 8308 /usr/sfw/lib/python2.3/heartbeat/crmd
> 6488 8356 /usr/sfw/lib/python2.3/heartbeat/crmd
>
>
> I tried to apply memory patches published in the past
> but they all seem to be already incorporated.
>
> _______________________________________________________
> Linux-HA-Dev: Linux-HA-Dev at lists.linux-ha.org
> http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> Home Page: http://linux-ha.org/
>
>
_______________________________________________________
Linux-HA-Dev: Linux-HA-Dev at lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/