On Thu, Sep 19, 2019 at 9:47 AM Guido Falsi <mad_at_madpilot.net> wrote: > On 19/09/19 18:04, Michael Gmelin wrote: > > > > What about gpart output of the pool drives? > > > > In general you would create zpools using gptids or gpt labels, not the > devices, so you’re independent of device numbering. The boot loader should > only be installed on drives that contain the boot pool (maybe you have old > boot loaders on data drives?). > > > > Actually the installer will at least some times use the adaX device to > create ZFS pools. At least it did for me when I recently rebuilt a > machine after a (multiple) disk crash. > > So it could not be Kurt fault if he has a pool with adaX devices in it. > > I installed the system on one disk and the installed used adaX to create > the pool. I added the second disk to the mirror a few days later. > > Now I have: > > pool: zroot > state: ONLINE > scan: resilvered 24.6G in 0 days 00:09:41 with 0 errors on Tue Sep 3 > 16:10:08 2019 > config: > > NAME STATE READ WRITE CKSUM > zroot ONLINE 0 0 0 > mirror-0 ONLINE 0 0 0 > ada0p4 ONLINE 0 0 0 > gpt/zfs1 ONLINE 0 0 0 > > errors: No known data errors > > Now, in the case of a mirror (or a zraid) this could be fixed by > detaching adaX devices and reattaching them using the label. Disvantage > is the cluster will need to resilver, causing some degraded time and > extra disk load. > Boot off a LiveCD/USB stick. Export the pool. Import the pool with -d /dev/gpt and it will use the GPT labels instead. Reboot into the system, and it should continue to use the GPT labels. No resilvering required. -- Freddie Cash fjwcash_at_gmail.comReceived on Thu Sep 19 2019 - 14:54:17 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:41:21 UTC