Occurs when a node cannot join the VCS cluster.
If you see this error in your Solaris or Linux error log, the node cannot join the cluster because VCS cannot start I/O fencing:
CVMCluster:???:monitor:node - state: out of cluster
Use vxdg list to verify that the vxfencoordg disk group is enabled on all nodes:
vxdisk list
If the vxfencoordg disk group is not enabled, deport and re-import the group:
vxdg deport vxfencoorddg vxdg -t import vxfencoorddg
Use vxdisk list to verify that the vxfencoordg disks are online:
vxdisk list
If the vxfencoordg disks are not online, restart VCS on each node:
hastop -local /etc/init.d/vxfen stop /etc/init.d/vxfen start hastart
Use vxfenadm -g to verify reservations exist on the vxfencoordg disks:
vxfenadm -g
If the reservations do not exist, stop the VCS, use vxfenclearpre to clear the VCS reservations, and restart VCS:
hastop -local /etc/init.d/vxfen stop vxfenadm -g all -f /etc/vxfentab vxfenclearpre /etc/init.d/vxfen start hastart
If a new node joins an existing cluster, you may need to use gapconfig -x to reapply the group membership configuration:
gabconfig -x
If you see this message in your Solaris or Linux error log, the Veritas configuration daemon cannot start:
ERROR: IPC Failure: Configuration daemon is not accessible
Clear and restart the configuration daemon:
vxconfigd -k "vxiod set 10 "vxconfigd -m disable "vxdctl init "vxdctl initdmp "vxdctl enable
If you see this error in your VCS error log, the disk group initialization failed when you ran vxdg init:
Device sda cannot be added to a CDS disk group
By default, vxdg initializes disks as CDS. Rerun the vxdg command with cds=off:
vxdg init disk_group cds=off