Node failure

In an FPO deployment, each node has locally attached disks. When a node fails or has a connection problem with other nodes in a cluster, disks in this node become unavailable. Reboot a node to repair a hardware issue or patch the operating system kernel. Both these cases are node failures.

If auto recovery is enabled, that is, the restripeOnDiskFailure=yes parameter is set to yes, and a failed node is recovered within auto recovery wait time (check the details described in the Auto Recovery for Disk Failure section), auto recovery handles the node failure automatically by bringing up down disks and ensuring all data has the desired replication. If a node is not recovered within the auto recovery wait time, auto recovery migrates the data off the disks in the failed node to other disks in cluster.