We have a setup that uses the gluster clients on the nodes where the storage is located for a replicated file system. Is there a way to expand the replica count and not have to remount the volume on the other nodes or delete/re-create the volume? Currently we do the following: gluster peer probe $clusteraddr mount -t glusterfs $clusteraddr:$global_volume $MOUNT_POINT get some data.... umount $MOUNT_POINT echo 'y' | gluster volume stop $global_volume echo 'y' | gluster volume delete $global_volume gluster volume create $global_volume replica $numreplicas node1:back_store node2:/back_store node3:/backstore gluster volume start $global_volume mount -t glusterfs localhost:$global_volume $MOUNT_POINT find $MOUNT_POINT -print0 | xargs --null stat >/dev/null Then on the other nodes we have to umount $MOUNT_POINT mount -t glusterfs localhost:$global_volume $MOUNT_POINT -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://gluster.org/pipermail/gluster-users/attachments/20111031/87f6d167/attachment.htm>
