From: "David Harris" <[EMAIL PROTECTED]>
...
What I need to do is restore a superblock that will force it to view sdb
(the second scsi disk when both are in) as the up-to-date version, and have
it sync sda. I'll look into using "mkraid --force-resync" or mucking with
the superblock myself.
In the future, I think I'm going to keep superblock backups. Three for each
device: both in sync, a is correct and b is out of date, and b is correct
and a is out of date. This way I can restore the superblocks and force a
sync whichever way I want.
enclosed is a copy of mingo's comments for a similar situation.
> While on the subject of RAID0, I also wondered about a single
> superblock for the autostart information. Since I've experienced
> superblock corruption on a Solaris system, and recovered from it by
> utilizing one of the superblock copies, could this be a single point
> failure location? Can the RAID code stash a superblock copy
> elsewhere, or is it easily recreated from the /etc/raid0.conf file?
the RAID architecture deals with this idssue and has several mechanizms to
protect against superblock lossage:
- every disk in the array has it's own superblock. The whole RAID
state can be restored from a single superblock.
- there is a per-superblock 32-bit checksum to detect corruption
or partial writes. (the superblock is multiple sectors, so it
might happen that an update gets executed only partially)
- there is an 'last update timestamp'. If a disk is offline
temporarily and the array is reconfigured, then the timestamp
helps us to not use the 'old' superblock.
[ - if you maintain /etc/raidtab properly (so that it really
represents the array) it might be used to recreate the array _if
there is no failed disk in the array_, without data loss.]
(actually, an invalid checksum doesnt mean we do not consider the
superblock, it just means that we consider it to be a 'very old'
superblock. So if everything else fails we try even such a superblock)
and as a generic rule, we never write superblocks back if an array has a
fatal failure. (this policy makes no difference to a completely failed
array, because a lost array is a lost array. But it might save us if there
is some other hardware component that has failed, after system restart we
might have a working array)
if there is anything else that should be added as a protection mechanizm,
let me know. we can never be paranoid enough about RAID arrays ;)
-- mingo
