Through sheer unimaginable good luck and nothing else, the resilvered
disk has decided to play ball and the pool has come back online..,
It's currently performing a scrub, this time, I'll make sure I've got
spare disks and backups aplenty.
BTW, What's the sparc version of OpenIndiania like? Anyone on the list
using it on a daily basis?
Thanks again,
Peter.
On 12/05/2013 1:17 AM, Peter Firmstone wrote:
Stupid things I did recently, I made a bad situation much worse:
1. Haven't backed up recently (busy put it off)
2. Discovered 4 bad blocks on one disk in a three disk raidz, that
had two disks running and one unavailable (later realised that
scsi bus resets caused by the bad blocks had made it unavailable).
I used format to repair the bad blocks.
3. Not thinking, I replaced the unavailable disk with a new disk
(should have just added a spare).
4. The replacement disk resilvered.
5. When resilvering was complete I shutdown and rebooted.
6. The kernel panicked after rebooting.
7. I exported the pool so I could reboot.
8. Tried to import the pool, with the same result; kernel panic.
9. It appears as though the replacement disk resilvering was
unsuccessful, so now I only have two disks.
I'm still running Solaris 10 sparc, haven't upgraded yet. I lack some
of the fancy zpool command line options that roll back transactions -F.
If I could roll back to an earlier zfs transaction, is it possible to
include the disk I removed and import the pool?
Any ideas?
Thanks in advance,
Peter.
_______________________________________________
OpenIndiana-discuss mailing list
[email protected]
http://openindiana.org/mailman/listinfo/openindiana-discuss
