I have two sets of 3 disks striped & mirrored - been running fine for about 2 months. Yesterday I saw the console report that it took one half of the mirror offline because a disk reported an internal failure. Data was written out since that event. Figuring it may come back up with a reboot, I did so. Seems to have been a bad choice as the data now seems to have reverted to the pre disk error versions. It has Rational Clear Case on it. CCase guys here don't thing it's a CCase event to make it revert.
My questions are: What does raid do in this case? Faced with what seems to be two versions of data, I would have thought it would have chosen the newer data version as the real & replicated it to the other half. I was thinking I should have mounted it up on the good half of the mirror only, destroy/rebuild the bad strip half, remount it like normal. Can someone explain what the underlying operations are for a mirror for writing, reading and error recovery like this? Thanks. -- toby _______________________________________________ Seawolf-list mailing list [EMAIL PROTECTED] https://listman.redhat.com/mailman/listinfo/seawolf-list
