I am unable to unlock one of my pools after a reboot recently, the other pool has no issues. When going through the UI, everything acts normal until I submit the passphrase, then that screen hangs at "Starting Job...". If I jump to a different tab, I see a `pool.dataset.unlock` task at 0% and the pools page seems to act like it's unlocked (Padlock is unlocked icon, options to lock pool again), but no data is actually there and none of the alerts for shares that use the encrypted pool clear.
This pool is made up of 2 Raid-Z1 Vdevs with a L2ARC and Spare configured. It has about 7 data sets on it, all are on the default settings and inherit encryption. System is Xeon E3-1220 V2 with 32GB of ECC RAM and 2 LSI HBAs. All the disks in question are on one HBA, but all disks report as Online in ZFS. Looking at disk I/O, there's not much going on, though some drives have a pretty constant read of 0.2 MB/s
What's interesting is just before this I was attempting to clean one of the data sets, this particular one has dedupe on, and an `rm` command hung for several minutes in a D+ state. I rebooted the node to clear that stuck command and was going to just remove that entire dataset when the pool ran into this issue.
I'm also totally fine losing this dataset, as I somehow suspect it's the problem. Unlocking and de-selecting "unlock children" yields the same results.
This pool is made up of 2 Raid-Z1 Vdevs with a L2ARC and Spare configured. It has about 7 data sets on it, all are on the default settings and inherit encryption. System is Xeon E3-1220 V2 with 32GB of ECC RAM and 2 LSI HBAs. All the disks in question are on one HBA, but all disks report as Online in ZFS. Looking at disk I/O, there's not much going on, though some drives have a pretty constant read of 0.2 MB/s
What's interesting is just before this I was attempting to clean one of the data sets, this particular one has dedupe on, and an `rm` command hung for several minutes in a D+ state. I rebooted the node to clear that stuck command and was going to just remove that entire dataset when the pool ran into this issue.
I'm also totally fine losing this dataset, as I somehow suspect it's the problem. Unlocking and de-selecting "unlock children" yields the same results.
