[Linux-cluster] Filesystem repair on a Clustered-NFS Server

[Jonathan Barber]

On 22 November 2010 18:12, Randy Zagar <zagar at arlut.utexas.edu> wrote:
>
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Assume for a moment that you have an EL5 NFS Cluster.
>
> You have a clustered service "nfs-server1" that has an IP resource and
> several "child" filesystems tied to it so that the filesystems move
> when the IP moves through the cluster.  Each of these child
> filesystems has an nfs-export resource associated with it, so they are
> nfs-exported automatically.
>
> How do you repair one of those filesystems without taking down the
> entire service?  You can't unmount it while it's exported, and if you
> try to unexport/unmount manually the cluster restarts it for you...

Freeze/unfreeze should do it:
http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/6/html/Cluster_Administration/s2-admin-manage-ha-services-operations-cli-CA.html

> Do you have to remove the failed filesystem from the service?  Or can
> you "disable" one resource and leave the others in peace?
>
> This is not entirely a rhetorical question for me,
>
> - -RZ
-- 
Jonathan Barber <jonathan.barber at gmail.com>