Jakob Østergaard wrote: > > Morale: Use RAID but Keep Good Backups. > ----------------- > > Unlike RAID, a backup will let you recover (minus one days work) from both > administrative/user disasters (rm -rf, mke2fs, dd, ...), and the physical ones. Nobody ever denied that. My point was that extra redundancy is a good thing to have while swapping one broken disk. A backup will recover from total disaster, but on for instance a large database system even losing just half a day of work can cost one and a half day to recover from. Half a day to restore the backup and bring the system up, half a day for the users to moan and groan and figure out how much work they lost exactly (digging related paperwork out of the wastebins) and another half day for them to redo the lost work. Not only can this be expensive, it causes general irritation too, which is never a good thing. There are many levels of redundancy, but RAID6 is an interesting one, with extra redundancy at relatively little extra cost compared to other options like mirroring and RAID5+5. It's always a tradeoff between costs and redundancy. Given unlimited money, you can build a system that will survive anything but a global nuclear war, but then you'd have other things to worry about than your database. As Dale said though, the parity scheme is probably not as easy as I thought. I didn't really put much thought in it yet. Marcel - To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
