On Sat, 2007-11-24 at 22:53 -0700, Pawel Osiczko wrote: > Hello, > > I'm having problems coming up with a solution which would > allow for a clustered RAID1 storage with 2 disks living > on a shared bus attached to 2 HA cluster nodes. So, essentially, > I need shared storage for the clustered nodes with caveat that RAID > should be capable of being failed over from one node to another. > Is such an animal possible? Seems like OCFSv2 and EVMS are not capable > of that. Redhat CLVM/GFS holds some promise, but I haven't had a chance > to evaluate it yet. Any ideas?
I am not sure I have understood exactly what you want or need here. Why not install an external RAID chassis with a dual channel controller? SCSI or fibrechannel, and connect each channel to one of the HA cluster nodes? You can implement any RAID type you like on the external array (given of course that the controller supports it). If you'tr're looking for a clustered filesystem which will cope with HA type failover, then I back up Michael's recommendation to look at Lustre. There is documentation on the Lustre site for configuring Lustre to work with Linux-HA. We have this working. (Actually, there is a small patch to the HA script which does the disk mounting, specifically the exclude list of filesystems which are not fsck'd when mounted. As lustre is another filesystem type it has to be added.) _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
