[Pacemaker] umount filesystem the node reboot

[jiaju liu]

I use lustre in my HA-cluster,sometimes when I umount ost, the node will reboot.an 20 09:23:10 oss1 kernel: Lustre: server umount testTwo-OST0003 completeJan 20 10:28:56 oss1 kernel: Lustre: server umount testTwo-OST0000 completeJan 20 10:28:58 oss1 kernel: exit dynlocks cacheJan 20 10:29:11 oss1 kernel: audit(1295490551.122:25403): audit_pid=0 old=2787 by auid=4294967295Jan 20 10:29:11 oss1 pcscd: pcscdaemon.c:572:signal_trap() Preparing for suicideJan 20 10:29:11 oss1 pcscd: hotplug_libusb.c:376:HPRescanUsbBus() Hotplug stoppedJan 20 10:29:12 oss1 pcscd: readerfactory.c:1379:RFCleanupReaders() entering cleaning functionJan 20 10:29:12 oss1 pcscd: pcscdaemon.c:532:at_exit() cleaning /var/runJan 20 10:29:12 oss1 kernel: Kernel logging (proc) stopped.Jan 20 10:29:12 oss1 kernel: Kernel log daemon terminating.Jan 20 10:29:13 oss1 exiting on signal 15
I think this may be the stonith resource reboot it?anyone else meet the same problem?I use single ring, I think the question is when umount several disks, the communication between nodes  blocks. redundant communication is helpful?


      
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://oss.clusterlabs.org/pipermail/pacemaker/attachments/20110120/506751d1/attachment-0001.htm>