[Gluster-users] What is the right way to bring down a Glusterfs server for maintenance?

[Strahil]

It's a replica 3 , right.

In my ovirt - I just run poweroff, but you can stop the glusterd.service and run this script:
/usr/share/glusterfs/scripts/stop-all-gluster-processes.sh

I think plain shutdown is enough.

After powering off , you can force a heal.

I have created  my own functions in .bashrc - 1 for full heal (using sharding here) and the second to check the status:


gluster-heal-all()
{
        for i in $(gluster volume list)
                do
                        gluster volume heal $i full
                done
}

gluster-heal-info()
{
        for i in $(gluster volume list)
                do
                        gluster volume heal $i info summary
                        echo;echo
                        sleep 2
                done
}

P.S.: Even if you don't trigger the heal - gluster has internal mechanism to do it for you.

Best Regards,
Strahil NikolovOn Jul 3, 2019 22:39, Carl Sirotic <csirotic at evoqarchitecture.com> wrote:
>
> I have a replica 3 cluster, 3 nodes with bricks and 2 "client" nodes, 
> that run the VMs through a mount of the data on the bricks. 
>
> Now, one of the bricks need maintenance and I will need to shut it down 
> for about 15 minutes. 
>
> I didn't find any information on what I am suposed to do. 
>
> If I get this right, I am suposed to remove the brick completely from 
> the cluster and add them again when the maintenance is finished ? 
>
>
> Carl 
>
> _______________________________________________ 
> Gluster-users mailing list 
> Gluster-users at gluster.org 
> https://lists.gluster.org/mailman/listinfo/gluster-users