[torqueusers] Problem with mlockall in resmom on aix: mom's die
with out of memory condition.
Michael Marti
michael.marti at ist.utl.pt
Mon Mar 30 18:33:01 MDT 2009
Dear All
pbs_mom of torque-2.3.6 breaks on aix:
In the file src/resmom/mom_main.c on line 7395 mlockall is called to
keep the OS from swapping resmom. It turns out that this call of
mlockall causes the memory consumption of the pbs_mom process to jump
from about 1MB to more that 250 MB.
A quick fix is to replace line 7386 which reads
#ifdef _POSIX_MEMLOCK
with
#if defined(_POSIX_MEMLOCK) && !defined(_AIX)
This solves the problem for us.
Would be nice if this gets fixed in a future version of torque.
Remains the question why mlockall behaves that way on aix.
uname -a on a node: AIX r1blade066 3 5 00003222D100
Best regards,
Michael Marti
--
----------------------------------------------------------------------------
Michael Marti
Instituto Superior Técnico
Instituto de Plasmas e Fusão Nuclear
Complexo Interdisciplinar
Av. Rovisco Pais
1049-001 Lisboa
Portugal
Tel: +351 218 419 379
Fax: +351 218 464 455
Mobile: +351 968 434 327
----------------------------------------------------------------------------
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.supercluster.org/pipermail/torqueusers/attachments/20090331/d090dd3f/attachment.html
More information about the torqueusers
mailing list