Ensure that node name and the service adapter label are the same on each node in the cluster

or

Alias the node name to the service adapter label in the /etc/hosts file.

5.4.5Cross Mounted NFS File Systems and the Network Lock Manager

If an NFS client application uses the Network Lock Manager, there are additional considerations to ensure a successful failover. Consider the following scenario: Node A has a file system mounted locally and exported for use by clients. Node B is an NFS client and mounts the exported file system for local use by an application that issues lock requests using the flock() system call. Node A fails. Node B then attempts to unmount the NFS mounted file system, mount it as a local file system, and export it for client use. However, the unmount fails because of outstanding lock requests against the file system.

Adding the following lines to the cl_deactivate_nfs script will clear outstanding locks against the failed node and will allow the file system to be unmounted. However, it will result in the loss of all locks. Consider your configuration carefully. If you have non-cluster related NFS file systems where losing locks would be unacceptable, you may need to take appropriate steps before using this addition to the cl_deactivate_nfs script.

Add the code below between the following two lines (three places):

128 IBM Certification Study Guide AIX HACMP

Page 146
Image 146
IBM SG24-5131-00 manual Cross Mounted NFS File Systems and the Network Lock Manager