Hi guys I have 4 node cluster with OCFS2.>From time to time redhat fired : oom-killer to kill process with high amoutof memory ( for example tomcat) My friends suppose that ocfs2 consume memory. So I have question: How check how many memory is occupied by ocfs processes? Is it a low memory? regards -------------- next part -------------- An HTML attachment was scrubbed... URL: http://oss.oracle.com/pipermail/ocfs2-users/attachments/20061031/6ef710b0/attachment.html
John Peeken, Enterprise Linux Support
2006-Oct-31 07:25 UTC
[Ocfs2-users] Ocfs2 and low memory
Rafal, What version of Red Hat are you using? cat /proc/meminfo will show us how memory (Low and High) is doing, top will show how much a process is using of virtual memory How much memory? What kernel is being used? Regards, John Peeken Rafal Maliszewski wrote:> Hi guys > > I have 4 node cluster with OCFS2. > From time to time redhat fired : oom-killer to kill process with high > amout of memory ( for example tomcat) > > My friends suppose that ocfs2 consume memory. > > So I have question: > How check how many memory is occupied by ocfs processes? > Is it a low memory? > > regards > > >------------------------------------------------------------------------ > >_______________________________________________ >Ocfs2-users mailing list >Ocfs2-users@oss.oracle.com >http://oss.oracle.com/mailman/listinfo/ocfs2-users > >-------------- next part -------------- An HTML attachment was scrubbed... URL: http://oss.oracle.com/pipermail/ocfs2-users/attachments/20061031/f1517a4d/attachment.html
John Peeken, Enterprise Linux Support
2006-Oct-31 08:20 UTC
[Ocfs2-users] Ocfs2 and low memory
Rafal, See if you have vm.lower_zone_protection set? # cat /proc/sys/vm/lower_zone_protection If not set, then set to 100 and viola all your low mem issues should vanish. You can use top -p <pid> , for the pid for ocfs2 and see how it is using memory But, with RHEL4 to eliminate any source of an OOM - set vm.lower_zone_protection=100 in /etc/sysctl.conf Regards, John Peeken Rafal Maliszewski wrote:> Redhat 4ES update 3 - kernel 2.6.9.34-smp > > Something is eating low memory, not highmem. > Howto check that there is not ocfs2?? > > > > On 10/31/06, John Peeken, Enterprise Linux Support > <john.peeken@oracle.com <mailto:john.peeken@oracle.com>> wrote: > > Rafal, > What version of Red Hat are you using? > cat /proc/meminfo will show us how memory (Low and High) is doing, > top will show how much a process is using of virtual memory > How much memory? What kernel is being used? > Regards, > John Peeken > > > > Rafal Maliszewski wrote: > >> Hi guys >> >> I have 4 node cluster with OCFS2. >> From time to time redhat fired : oom-killer to kill process with >> high amout of memory ( for example tomcat) >> >> My friends suppose that ocfs2 consume memory. >> >> So I have question: >> How check how many memory is occupied by ocfs processes? >> Is it a low memory? >> >> regards >> >> >>------------------------------------------------------------------------ >> >>_______________________________________________ >>Ocfs2-users mailing list >>Ocfs2-users@oss.oracle.com >> <mailto:Ocfs2-users@oss.oracle.com> >>http://oss.oracle.com/mailman/listinfo/ocfs2-users >> >> >-------------- next part -------------- An HTML attachment was scrubbed... URL: http://oss.oracle.com/pipermail/ocfs2-users/attachments/20061031/a9c1d84a/attachment-0001.html
To monitor ocfs2 memory usage, do: # cat /proc/slabinfo | egrep 'ocfs|dlm|size-256 |size-32 ' ocfs2_lock 16 226 16 226 1 : tunables 120 60 0 : slabdata 1 1 0 ocfs2_inode_cache 22 24 1152 3 1 : tunables 24 12 0 : slabdata 8 8 0 ocfs2_uptodate 28 119 32 119 1 : tunables 120 60 0 : slabdata 1 1 0 ocfs2_em_ent 10 183 64 61 1 : tunables 120 60 0 : slabdata 3 3 0 dlmfs_inode_cache 1 5 768 5 1 : tunables 54 27 0 : slabdata 1 1 0 dlm_mle_cache 0 0 384 10 1 : tunables 54 27 0 : slabdata 0 0 0 size-256 40245 40245 256 15 1 : tunables 120 60 0 : slabdata 2683 2683 0 size-32 41650 41650 32 119 1 : tunables 120 60 0 : slabdata 350 350 0 # cat /proc/fs/ocfs2_dlm/*/stat local=26, remote=0, unknown=0 local=39963, remote=6, unknown=0 size-256/32 are generic slab caches but are also used by ocfs2dlm. The ocfs2dlm impact on it can be detected by the second cat which lists the number of locally mastered (local) locks which are currently not freed until the volume is umounted. (The patch-fix is being tested.) Rafal Maliszewski wrote:> Hi guys > > I have 4 node cluster with OCFS2. > From time to time redhat fired : oom-killer to kill process with high > amout of memory ( for example tomcat) > > My friends suppose that ocfs2 consume memory. > > So I have question: > How check how many memory is occupied by ocfs processes? > Is it a low memory? > > regards > > > ------------------------------------------------------------------------ > > _______________________________________________ > Ocfs2-users mailing list > Ocfs2-users@oss.oracle.com > http://oss.oracle.com/mailman/listinfo/ocfs2-users >