Peter Santos
2008-Dec-19 17:43 UTC
[Ocfs2-users] Which version of ocfs2/tools is stable on SLES10 SP2
Guys, Quick question? We have had ocfs2 1.2 running for a few years now in our SLES9 SP3 environments, and these are used just fine for offloading database backups and archivelogs from Oracle's ASM (10.2.0.3). Last week we installed SLES10 SP2 and ocfs2 on a 2 node RAC cluster using Oracle 11g. The 2 nodes stay up just fine with the Oracle clusterware, and even with the ocfs2 mounted on both nodes. However when we try to scp data onto either of these machines from the outside, somewhere in the middle of the scp, one of the nodes locks up and crashes. We monitered the interconnect and the reason the machine is crashing is because at some point during the copy, the interconnect goes dead and no more packets are able to pass through the interconnect. When doing the scp, it doesn't matter if we are scp'ing to the ocfs2 partition or to a local partion on a completely separate disk. ..either way will cause usually the first node to lock up. We are using ocfs2 1.4.0 and not sure if this is really stable. We noticed the ocfs2_hb_ctl utility generated some errors during install and it seems like no matter what we do, we can't get the O2CB_HEARTBEAT_THRESHOLD to change from its default value. Even if we update the O2CB_HEARTBEAT_THRESHOLD value in the /etc/sysconfig/o2cb , the value from the output of /etc/init.d/o2cb status is always 31. Any experiences where 1.4.0 is unstable? Also where do I get 1.4.1 ? thanks. -peter
Sunil Mushran
2008-Dec-19 17:50 UTC
[Ocfs2-users] Which version of ocfs2/tools is stable on SLES10 SP2
Get the latest sles10 sp2 kernel. It includes 1.4.1-SLES. Peter Santos wrote:> Guys, > Quick question? > > We have had ocfs2 1.2 running for a few years now in our SLES9 SP3 > environments, and these are used > just fine for offloading database backups and archivelogs from Oracle's > ASM (10.2.0.3). > > Last week we installed SLES10 SP2 and ocfs2 on a 2 node RAC cluster > using Oracle 11g. > > The 2 nodes stay up just fine with the Oracle clusterware, and even > with the ocfs2 mounted > on both nodes. However when we try to scp data onto either of these > machines from the outside, somewhere > in the middle of the scp, one of the nodes locks up and crashes. We > monitered the interconnect and the > reason the machine is crashing is because at some point during the > copy, the interconnect goes dead and no > more packets are able to pass through the interconnect. > > When doing the scp, it doesn't matter if we are scp'ing to the ocfs2 > partition or to a local partion > on a completely separate disk. ..either way will cause usually the > first node to lock up. > > We are using ocfs2 1.4.0 and not sure if this is really stable. We > noticed the ocfs2_hb_ctl utility generated > some errors during install and it seems like no matter what we do, we > can't get the O2CB_HEARTBEAT_THRESHOLD to change > from its default value. Even if we update the O2CB_HEARTBEAT_THRESHOLD > value in the /etc/sysconfig/o2cb , the value > from the output of /etc/init.d/o2cb status is always 31. > > Any experiences where 1.4.0 is unstable? Also where do I get 1.4.1 ? > > thanks. > > -peter > > > _______________________________________________ > Ocfs2-users mailing list > Ocfs2-users at oss.oracle.com > http://oss.oracle.com/mailman/listinfo/ocfs2-users >