On 10/26/07, Jake Conk <[EMAIL PROTECTED]> wrote:
If the filesystem is screwed up then shouldn't the raid just ignore it
and run on 1 disk until I fix the problem? That seems like the
logical thing it should do unless all my mirrors of /var are messed
up.
No, raid doesn't do that.
Let's assume we have a raid1 setup with two twin disks, no spares.
Imagine it as a floor (your filesystem) substained by two columns (your
disks): if a column collapses the other one keeps the floor up, but if
the floor has holes your forniture will fall down, regardless columns'
health.
RAID just cares about disks' health, not filesystem's health.
And lastly, is it possible in the worst case scenario if one of my
disks is completely fsck'ed up is it possible to run the system on 1
of the raid 1 disks until a second comes?
If the *disk* fails raid takes care of it automagically. If for some
reasons the filesystem fails then you might have to fix it by hand.
f.
