Quick questions, quick answers.
Matthew Economou wrote:
> 1. Where is Martin Bene's failed disk patch?
http://linuxhq.jimpick.com/lnxlists/linux-raid/lr_9904_04/msg00030.html
> 2. I am using the exact same disk configuration and raidtab. Do I
> need to do anything special to verify correct ordering, etc?
If you haven't swapped driver order/lun since you installed, then no. but
make absolutely sure, otherwise you'll get a destroyed file system (actually
not quite, but it's still bad)
> 3. Do I really need the failed disk patch? If the event counters
> (14 and 15) are to be believed, the two "failed" disks are
> pretty close to each other.
Yes. From observation, it would seem that the event counters are
startup/shutdown counters of some type... they aren't like blocks written
counters. Most of the time when it really needs to sync the counters will
only be one apart.
Your drives were going through re-sync. Even though in theory the data was
the same from the last time they were in sync, I don't think I'd assume that
the behavior of raid was that it would not leave any trash, partially
unwritten data or whatever lying around while it's doing it's resync. If
you assume wrong, you could hose your whole array. If you cause a resync we
know you'll be OK.
> 4. What about "mkraid --force-resync"?
I don't really know what this does, and it's not in the man page. I really
doubt it does what you need to do, as if it did I expect someone would have
pointed that out to us. If Igno didn't suggest it, I highly doubt it's up
to the task. Just by name alone, I suspect that it makes your parity drive
re-sync when you're not otherwise in degraded mode.
> I may try the resync, because it seems to be really close to what you
> (and Ingo Molnar (in private email) and Piete Brooks) suggest. But I
> *will* wait for your reply. :)
Good luck
Tom
