Gremlins in my machine have deleted most of the pool metadata and I
need to know if it is possible to recover.
Setup is 8 disks in an external enclosure configured as raidz2
attached to a macpro running openindiana under vmware fusion.
This has worked fine for a number of months with the only caveat that
whenever the Mac is rebooted you have to rebuild the virtual disk
If you don't then vmware complains of missing disks and won't start
openindiana or zfs complains of corruption. When they are rebuilt and
properly named then there
are no problems.
I updated fusion so I had to reboot and when I did my script failed
because vmware had moved their utility so I got the expected
corruption message from zfs.
However this time it did not go away when the disk descriptions etc
were properly done.
On the Mac diskutility now reports some disks as unlabeled and
unformatted and some with just a GUID label. Before they were all efi
zfs. Maybe this is causing zfs to look in the wrong places.
Looking at the disks with dd etc some have the labels and metadata
basically all nulls some have garbage and some appear to have good
Is it possible to have that good info copied to other disks into the
right places to allow the pool to be recovered ?
Any ideas and which part of that system actually deleted the data?
zfs-discuss mailing list