If an HCP node fails, NFS mounts that target the failed node lose their connections to the namespace. To recover from a node failure, unmount the namespace at the current mount point. Then take one of these actions:
- Mount the namespace on a different node. You can do this by specifying either the domain name of the HCP system or a different node IP address in the mount command. If you specify a domain name, HCP automatically selects a node from among the healthy ones.
- When the failed node becomes available again, remount the namespace on that node.