Yvette, You can: - either use GFS2 with concurrent access to your filesystem from both nodes, as this is a cluster filesystem - or use ext3/XFS as a failover filesystem - mounted by no more than one of the cluster nodes at any time. Either of the two approaches will have HA characteristics without any need for NFS. Regards, Chris Jankowski -----Original Message----- From: linux-cluster-bounces@xxxxxxxxxx [mailto:linux-cluster-bounces@xxxxxxxxxx] On Behalf Of yvette hirth Sent: Monday, 29 November 2010 14:46 To: linux clustering Subject: new cluster defined and storage added - what's next? hi, ok i've been lurking here for a while studying "the whole cluster thing", and i've made this progress: two ILO2-based HP servers running CentOS 5.5 with fibre connections to a switch; one promise fibre array-based raid1 drive of 4TB (4x 2TB drives); and Cluster Storage and Clustering packages installed. using luci running on srv0 i've defined the cluster, added two nodes, their fencing (IPMI Lan), and everything looks good. on srv0 i changed the "wrk" raid1 storage to "Clustered: True", and sure enough it shows up on srv1 as "Clustered: True". so i assUme i've done everything right thus far... since i'm aiming for a primary:primary cluster (HA), i should change that cluster.conf parm for that, yes? if i want to be able to access this "wrk" drive from a non-clustered server via NFS, what's next? q: the "wrk" array uses xfs (setup before the cluster). it's ok if it gets levelled. do i need to use GFS2 or ZFS or some other filesystem format, or can i continue to use xfs? q: i'm assUming that i need to make NFS a managed service if i want it to be HA, yes? q: anything i've missed? i'm amazed that this has been this easy thus far - kudos to all of you! yvette -- Linux-cluster mailing list Linux-cluster@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/linux-cluster -- Linux-cluster mailing list Linux-cluster@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/linux-cluster
