Hello all,
I am attempting to do something that I have done many times before, but I am unable to now. I am replacing a disk in my array in order to grow the array.
Current setup is FreeNAS-9.2.0-RELEASE-x64. Array (tank0) has 2 x 500GB and 2 x 1TB drives. I am trying to replace one of the 500GB drives with another 1TB drive. However a bunch has changes since I have last replaced a drive to go the array.
Change:
1) I believe I upgraded (GUI) from 8.2 to 8.3 (x86)
2) Updated ZFS of this pool to v28
3) Changes computer hardware that this FreeNAS installation runs
4) Updated from 8.3 (x86) to 9.2 (x86)
5) Updated from 9.2 (x86) to 9.2 (x64)
Upgrade process is the same (this is for ada0)
1) Login via SSH
2) Verify the status of the pool and last scrub (done weekly)
3) Bring the drive to be replaced offline (zpool offline tank0 gptid/<UUID>)
4) Shutdown FreeNAS and replace physical drives
5) Via command line, verify new ada0 is the drive I just inserted
6) Use gpart to destroy, create, and partition with freebsd-zfs
7) Replace device in zpool (zpool replace tank0 ... ...)
8) Detach old device from the pool (zpool detatch tank0 ...)
My problem comes at step 7. Now with 9.2, device names are no longer used and gptid's are in their place. I have been using in the zpool command "gptid/uuid" in place of where I historically have used ada0p1.
My result:
Upon issuing the replace device command, the zpool process enters a "D" sleeping state and the command prompt never comes back to me. The pool is useless, and I eventually have to reboot (forcing power off and power on) to get response back to the pool. Other aspects of the FreeNAS environment seem fine as long as I don't do something with the pool (such as look at reporting, shell, process list, etc).
I have tried various ways of specifing the "zpool replace" command with /dev/gptid/<uuid>, or trying old school device names such as /dev/ada0p1, or even trying to replace via the GUI. They all have the same result. Process goes to sleep and the pool sits unusable.
I can, however, put the old drive back in the system, and bring it back online. It will resilver and scrub just fine.
Any thoughts on how I can get this new drive to replace the old drive? Once I can get this to work I have another 1TB drive to replace the last 500GB drive.
Thanks for your help.
I am attempting to do something that I have done many times before, but I am unable to now. I am replacing a disk in my array in order to grow the array.
Current setup is FreeNAS-9.2.0-RELEASE-x64. Array (tank0) has 2 x 500GB and 2 x 1TB drives. I am trying to replace one of the 500GB drives with another 1TB drive. However a bunch has changes since I have last replaced a drive to go the array.
Change:
1) I believe I upgraded (GUI) from 8.2 to 8.3 (x86)
2) Updated ZFS of this pool to v28
3) Changes computer hardware that this FreeNAS installation runs
4) Updated from 8.3 (x86) to 9.2 (x86)
5) Updated from 9.2 (x86) to 9.2 (x64)
Upgrade process is the same (this is for ada0)
1) Login via SSH
2) Verify the status of the pool and last scrub (done weekly)
3) Bring the drive to be replaced offline (zpool offline tank0 gptid/<UUID>)
4) Shutdown FreeNAS and replace physical drives
5) Via command line, verify new ada0 is the drive I just inserted
6) Use gpart to destroy, create, and partition with freebsd-zfs
7) Replace device in zpool (zpool replace tank0 ... ...)
8) Detach old device from the pool (zpool detatch tank0 ...)
My problem comes at step 7. Now with 9.2, device names are no longer used and gptid's are in their place. I have been using in the zpool command "gptid/uuid" in place of where I historically have used ada0p1.
My result:
Upon issuing the replace device command, the zpool process enters a "D" sleeping state and the command prompt never comes back to me. The pool is useless, and I eventually have to reboot (forcing power off and power on) to get response back to the pool. Other aspects of the FreeNAS environment seem fine as long as I don't do something with the pool (such as look at reporting, shell, process list, etc).
I have tried various ways of specifing the "zpool replace" command with /dev/gptid/<uuid>, or trying old school device names such as /dev/ada0p1, or even trying to replace via the GUI. They all have the same result. Process goes to sleep and the pool sits unusable.
I can, however, put the old drive back in the system, and bring it back online. It will resilver and scrub just fine.
Any thoughts on how I can get this new drive to replace the old drive? Once I can get this to work I have another 1TB drive to replace the last 500GB drive.
Thanks for your help.
