Re: Allocation failures
David Brodbeck <[email protected]> Mon, 4 Jun 2007 09:18:58 -0700
| Newsgroups | gmane.linux.cluster.openmosix.general |
|---|---|
| Message-ID | <[email protected]> |
It looks like the kernel was compiled with 4 GB support, so that should be OK. I wonder if enabling HIGHMEM I/O Support would help? I'm suspicious I'm running out of space in the low memory area, so maybe shifting the I/O buffers to high memory would mitigate the problem. On Jun 4, 2007, at 6:23 AM, Jürgen Knödlseder wrote: > What memory option did you use to compile the kernel? I was told that > 64 GB HIMEM is not supported, so I'm now running with the 4 GB option > (which in fact does not allow me to use the full memory of my > machines :-( Since I used the 4 GB option I did not encounter any > problems anymore (I has stability problems with 64 GB ...) > > Jürgen > > Le 4 juin 07 à 11:54, Ralf Oelschlaegel a écrit : > >> David Brodbeck schrieb: >>> I'm managing an OpenMosix cluster, running kernel 2.4.26-om1 on >>> Debian Sarge. After the head node has been up for a couple of weeks >>> I start seeing a lot of allocation failure messages in the kernel >>> logs. The situation eventually deteriorates to the point where the >>> machine is unusable -- all attempts to launch processes end with >>> fork >>> reporting allocation failures. The problem does not appear to be >>> memory exhaustion; the machine has 4 GiB of RAM and never uses more >>> than 100 megabytes or so of swap. >>> >>> Has anyone else seen this problem? Is this the infamous Linux 2.4.x >>> 'memory fragmentation' issue rearing its ugly head? Is there a fix? >> we use the same kernel on SuSE8.2 system and we have the same >> problem. After one >> month I have to reboot the master node and I think the problem is >> memory >> fragmantation (see memory/rawmemory on master by mosmon). >> Our "workaround" is checkpoint/restart for long running jobs. >> >> Ralf >> >> >> --------------------------------------------------------------------- >> - >> --- >> This SF.net email is sponsored by DB2 Express >> Download DB2 Express C - the FREE version of DB2 express and take >> control of your XML. No limits. Just data. Click to get it now. >> http://sourceforge.net/powerbar/db2/ >> _______________________________________________ >> openMosix-general mailing list >> openMosix-general-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org >> https://lists.sourceforge.net/lists/listinfo/openmosix-general > > > ---------------------------------------------------------------------- > --- > This SF.net email is sponsored by DB2 Express > Download DB2 Express C - the FREE version of DB2 express and take > control of your XML. No limits. Just data. Click to get it now. > http://sourceforge.net/powerbar/db2/ > _______________________________________________ > openMosix-general mailing list > openMosix-general-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org > https://lists.sourceforge.net/lists/listinfo/openmosix-general David Brodbeck Information Technology Specialist 3 Computational Linguistics ------------------------------------------------------------------------- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ _______________________________________________ openMosix-general mailing list openMosix-general-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org https://lists.sourceforge.net/lists/listinfo/openmosix-general