Glusterfs Node Offline repair method

Under a replicated volume, a single server node is offline and can be recovered by deleting all of the configuration files.

1, stop gluster related services, and kill all gluster related processes.

2, the configuration file only retains glusterd.info.

3, start the service.

Description

If the service fails to start because the Vol profile cannot be successfully loaded, you can delete the corresponding configuration file.

Because the node cannot communicate, it can be resolved by deleting the corresponding configuration file for peers.

A normal cluster can be forced to kick out the failed node through Gluster peer Detach force.

After the failed node is cleaned, the repaired node is added to the cluster through probe.

This article from "Members Doukua" blog, declined reprint!

Glusterfs Node Offline repair method