In theory recent versions of Gluster could possibly be used in such a configuration but its a bit adhoc.
Essentially you could configure two nodes as replicas with no striping on one in each site and put the bricks on an OCFS2, GFS, or VCFS filesystem and due to the auto self heal feature you could possibly use it in an active/hot spare scenario but it would be delicate to say the least.
Gluster is meant to be a replacement for share disk clustering solutions and really eventually the way development is headed may be a viable network replacement for RAID. In theory with the proper automation it can be used to replace RAID 1+0 now but it requires 3rd party monitoring and automation to evacuate failed bricks in the case of a physical disk failure which is delayed at best and dubious at worst.
It is unusually to use Gluster with a clustered filesystem as it is now. I can see future possibilities and give some excellent detailed design specs and use cases to any one who wants to write the enhancements to the code but that is not where it is now.
-- Sent from my HP Pre3
On Jan 21, 2014 19:19, Florin Andrei <florin@xxxxxxxxxxxxxxx> wrote:
Consider two server clusters. Each cluster has its own local SAN, over
Fiber Channel. Within each cluster, all nodes have access to the same
OCFS2 volume, stored on the SAN. The two clusters are in separate
geographic regions, with a fast link in between.
Currently, the two clusters are kept more-or-less in sync via rsync. A
node in one cluster shoots off rsync to a node in the other cluster and
syncs up recent changes. It doesn't really work very well.
Is it possible to install GlusterFS on top of OCFS2 / SAN to keep the
two clusters in sync more reliably?
The operating system is Oracle Linux 5 (64 bit).
--
Florin Andrei
http://florin.myip.org/
_______________________________________________
Gluster-users mailing list
Gluster-users@xxxxxxxxxxx
http://supercolony.gluster.org/mailman/listinfo/gluster-users
_______________________________________________ Gluster-users mailing list Gluster-users@xxxxxxxxxxx http://supercolony.gluster.org/mailman/listinfo/gluster-users
