Paco

Personal Content Organiser

Home / Linux

Navigation

Menu

Search

Resolving a GlusterFS Server Failure

Check the state

gluster volume info all gluster peer status gluster volume status all


csn-002.stage.local is not available

Remove the bricks



Distributed
gluster volume remove-brick iso csn-002.stage.local:/gluster/iso/brick force

check
gluster volume info iso


Replicated

Note

Reduce the number of total replica by 1

gluster volume remove-brick data replica 2 csn-002.stage.local:/gluster/data/brick force


Detach the disconnected node


gluster peer detach csn-002.stage.local

check
gluster peer status


On the new node install with glusterfs deamon installed and started, create the LVM, mounts and the brick directories

On one of the healthy node



1. Add the new node to the peer
gluster peer probe csn-002.stage.local

check
gluster peer status


2. Add the Bricks

Replicated

Note

Increase the replica set number by 1

gluster volume add-brick data replica 3 csn-002.stage.local:/gluster/data/brick

check
gluster volume info data


Distributed
gluster volume add-brick iso csn-002.stage.local:/gluster/iso/brick


Thats it!

GlusterFS will self heal the volume and sync the data.


Source:
https://biscuit.ninja/blog/resolving-a-glusterfs-server-failure