Veritas Storage Foundation 5.1 SP1 for Oracle RAC Release Notes (5900-1513, April 2011)

Node join can lead to hang if an upgrade of the cluster protocol
version is in progress (2103567)
If you attempt to join a node to the cluster while Cluster Volume Manager (CVM)
is upgrading the cluster protocol version, the system may hang. This issue occurs
if the node is attempting to join the cluster after you issue the vxdctl upgrade
command to upgrade the CVM cluster.
Work-around:
Avoid joining a new node to the cluster until the CVM cluster upgrade is completed.
When master node loses access to complete storage, detached
sites remain in RECOVER state even after reattaching and
recovering the sites
In a campus cluster environment, if the master node loses access to complete
storage, all but one of the sites is detached and the DCO volumes may get detached
if the dgfailpolicy is set to dgdisable. If the detached sites are reattached and
recovered, the site still remains in RECOVER state. [1828142]
Workaround: Change the status of the site as described in the following procedure
to resolve the issue.
To change the status of the site
1
Log onto the CVM master node.
2
Reattach the detached sites:
# vxdg -g dg_name reattachsite site_name
The site remains in RECOVER state.
3
Restore DCO volumes by unpreparing and preparing the volumes.
Unprepare the volumes:
# vxsnap -g dg_name -f unprepare vol_name
Prepare the volumes:
# vxsnap -g dg_name prepare vol_name drl=on
45Storage Foundation for Oracle RAC Release Notes
Known issues