I’m setting up a 2 node cluster using GFS on a SAN and everything is ok until one node is forcibly shutdown. The shutdown is done for testing the failover process. Initially it the service and resources fails over on the other node but once node 1 shutdowns and tries to umount the FS the active node 2 FS suddenly hangs. here’s the output using the group_tool command.
type level name id state
fence 0 default 00010001 FAIL_START_WAIT
dlm 1 rgmanager 00020001 none
dlm 1 GFS 00040001 FAIL_ALL_STOPPED
dlm 1 clvmd 00050001 none
gfs 2 GFS 00030001 FAIL_ALL_STOPPED