All true, but I didn't want to go all the way to total redundancies,
etc. I thought, however, that there was a way of preventing the
duplication of FAT corruption, but I could be wrong.
Earl
Manuel Viet wrote:
>Le samedi 18 Juin 2005 14:20, Earl Dunbar a écrit :
>
>
>>SOFTWARE-based mirroring is NOT a good way to go, as your example
>>proves. Mirroring needs to be implemented at the controller level, or
>>better yet, dual controllers. Mirrored RAID-5 would be very good.
>>
>>
>
>That wouldn't help either in case of allocation table corruption. The solution
>is hardware raid (0 or 5) [dual controller is better but wouldn't help in
>case of motherboard failure, not much different from a single controller
>failure : there's still a single point of failure in the chain], backed up on
>good ol' magnetic tape on a day / week / month scheme basis (the day tape is
>kept localy, the week tape is kept at home, and the month tape is in a bank
>safe). THAT is pricey, but worth it. And don't forget the yearly regeneration
>of all mediums, to keep up with technology : I still have perfect, 30 years
>old, good 8" floppies, but if my last working drive gives up, they'll become
>mute on the spot, forever.
>
>
>
==============================================
List usage info: http://www.zuikoholic.com
List nannies: olympusadmin@xxxxxxxxxx
==============================================
|