I would like to know gpl filesytems for a two node failover cluster something like drbd but much more reliable and less time to sync
entire filesystem than drbd
What do you perceive as being unreliable about drbd? What part of the sync takes too long?
Most other failover schemes involve shared disk arrays. Until serial ATA can overcome the cable length problems parallel ATA has, that will mean a much more expensive SCSI array. In that case, failover is simple. Any journaled filesystem should work. You just have to cable two systems to the shared storage and set them up to only mount the array r/w, one at a time. Red Hat AS includes software and documentation on this. If you have the cash for the SCSI array, AS should be a no-brainer.
-- redhat-list mailing list unsubscribe mailto:[EMAIL PROTECTED] https://www.redhat.com/mailman/listinfo/redhat-list
