I need to setup a new infrastructure which provides high available shared storage between debian servers.
Our current solution is a "clustered DRBD NFS solution". We have setup 2 virtual file servers and mirror the shared storage with drbd. We run NFS-Server (active/passive) on each fileserver. Other server mount the storage/shared folder via NFS. To control failover we use pacemaker/corosync.
Is this approach is still a good solution, or are there meanwhile better open source solutions e.g. with GlusterFS, etc?