I have 2 hitachi drives, very similar models/serial numbers, etc. I also have 2 Toshiba drives.
Was running 8.3.1-Beta.
I had the hitachis plugged into my sata-3 ports on the biostar fxe motherboard, but one of the 2 identical toshiba drives, was being managed by the intel 1015 raid in passthrough (IT) mode.
Made 2 encrypted zpools, with 1 disk each. data1, data2. These pools have been running fine for months. I got bored and today decided to add a third encrypted volume (data3) on the Toshiba, then extended it as a 2 disk striped zpool to the other Toshiba (on the 1015) then detached/deleted the volume (i was hoping it would turn it into a mirror, not spanned). It failed to update in the web interface on the left panel, so i rebooted. It was still there in the volume manager on the left, but didnt show up in the volume manager in the main window. Then i tried clicking on volume management for it, and i got a http error. I rebooted again, still there, and also one of my other 1 disk zpools (data1) had an ERROR. So I took out the Toshibas to simplify things, then rebooted. The boot log shows that the OS read both hitachi drives twice on ada0+2, and ada1+3. Very strange. Another reboot, and taking out some hot swap drives to simplify the bootup and hopefully recover data - soon BOTH data1/data2 drives were ERROR, and unreadable. I tried to auto-import, but both drives were showing as 'data2', so data1 appeared to be lost, as if the OS had written data2 stripe info over the data1 stripe info. The USB it was burned on was also old and maybe not 100% functional, but this is just speculation.
Anyway long story short, I burned 8.3.1-RELEASE-p2 to a new USB and re-imported both volumes with the keyfiles and passphrase. I recovered all my data, no problems - whew. No need to restore from an out-of-date backup. Still.. buggy as hell. I been using freenas/GELI since the get go.. and It's always been tough.
Was running 8.3.1-Beta.
I had the hitachis plugged into my sata-3 ports on the biostar fxe motherboard, but one of the 2 identical toshiba drives, was being managed by the intel 1015 raid in passthrough (IT) mode.
Made 2 encrypted zpools, with 1 disk each. data1, data2. These pools have been running fine for months. I got bored and today decided to add a third encrypted volume (data3) on the Toshiba, then extended it as a 2 disk striped zpool to the other Toshiba (on the 1015) then detached/deleted the volume (i was hoping it would turn it into a mirror, not spanned). It failed to update in the web interface on the left panel, so i rebooted. It was still there in the volume manager on the left, but didnt show up in the volume manager in the main window. Then i tried clicking on volume management for it, and i got a http error. I rebooted again, still there, and also one of my other 1 disk zpools (data1) had an ERROR. So I took out the Toshibas to simplify things, then rebooted. The boot log shows that the OS read both hitachi drives twice on ada0+2, and ada1+3. Very strange. Another reboot, and taking out some hot swap drives to simplify the bootup and hopefully recover data - soon BOTH data1/data2 drives were ERROR, and unreadable. I tried to auto-import, but both drives were showing as 'data2', so data1 appeared to be lost, as if the OS had written data2 stripe info over the data1 stripe info. The USB it was burned on was also old and maybe not 100% functional, but this is just speculation.
Anyway long story short, I burned 8.3.1-RELEASE-p2 to a new USB and re-imported both volumes with the keyfiles and passphrase. I recovered all my data, no problems - whew. No need to restore from an out-of-date backup. Still.. buggy as hell. I been using freenas/GELI since the get go.. and It's always been tough.
