Hi guys, We are running a small compute cluster with 18 hosts, each has six 900GB SAS, we use two drives in RAID1 for OS. So there still have four free drives. I am planning to deploy a GlusterFS from the remaining SAS drives on each node. The storage will be used mainly as the scratch space of the running genome analysis (big data). Since the scratch space is for temporary processing data, so we want it to be fast and as big as possible. So any recommendation for configuring these drives to be used by Gluster? The RAID controller can do RAID0, RAID1, JBOD and just leave them as individual disks. Thanks Derrick -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20130301/c7e41b0c/attachment.html>