ht7261 at ot.x0.to
2018-Jul-01 07:53 UTC
zfs rootfs mount failure after upgrade to 11.2R from 11.1R-p10
Hello, zfs rootfs mount fails for me after an upgrade from 11.1R-p10 to the recent 11.2R. I upgraded my amd64 11.1R-p10 system to 11.2R using "freebsd-update upgrade". On the first reboot after the first "freebsd-update install", the system fails when the new kernel is about to mount its root filesystem. -------------------- : FreeBSD 11.2-RELEASE #0 r335510: Fri Jun 22 04:32:14 UTC 2018 root at releng2.nyi.freebsd.org:/usr/obj/usr/src/sys/GENERIC amd64 : : ukbd0 on uhub1 ukbd0: <ATEN Advance Tech Inc. USB DVI KVM, class 0/0, rev 1.10/1.00, addr 3> on usbus0 kbd2 at ukbd0 Mounting from zfs:zpool1/ROOT/default failed with error 6; retrying for 3 more seconds Mounting from zfs:zpool1/ROOT/default failed with error 6. Loader variables: vfs.root.mountfrom=zfs:zpool1/ROOT/default : : mountroot> ? List of GEOM managed disk devices: gpt/zfs0 msdosfs/EFI gpt/swap0 msdosfs/EFISYS gpt/EFI nvd0p3 nvd0p2 nvd0p1 cd0 nvd0 mountroot> -------------------- If I choose 'kernel.old' at the Beastie menu, the system boots the previous 11.1-p10 fine. (This was also updated from previous 11.1-pX by freebsd-update.) Below is the case for single user mode boot that I tried AFTER the above mentioned error. -------------------- : FreeBSD 11.1-RELEASE-p10 #0 Tue May 8 05:21:56 UTC 2018 root at amd64-builder.daemonology.net:/usr/obj/usr/src/sys/GENERIC amd64 : : ukbd0 on uhub1 ukbd0: <ATEN Advance Tech Inc. USB DVI KVM, class 0/0, rev 1.10/1.00, addr 3> on usbus0 kbd2 at ukbd0 Enter full pathname of shell or RETRUN for /bin/sh: Cannot read termcap database; using dumb terminal settings. # mount zpool1/ROOT/default on / (zfs, local, noatime, read-only, nfsv4acls) devfs on /dev (devfs, local, multilabel) # zpool status pool: zpool1 state: ONLINE scan: scrub repaired 0 in (..snip..) config: NAME STATE READ WRITE CKSUM zpool1 ONLINE 0 0 0 nvd0p3 ONLINE 0 0 0 errors: No known data errors # gpart show => 40 500118112 nvd0 GPT (238G) 40 1600 1 efi (800K) 1640 480 - free - (204K) 2048 134217728 2 freebsd-swap (64G) 134219776 365897728 3 freebsd-zfs (174G) 500117504 648 - free - (324K) -------------------- No hardware configuration nor BIOS/UEFI settings were touched in between the two cases. Were there anything that might have caused this problem between 11.1 and 11.2? This is an UEFI-only setup with NVMe SSD configured for root-on-zfs, zfs-only boot. and it is an 'Intel Core i5-7600T @ 2.00GHz' system on 'ASRock H270 Pro4' M/B with 64GB RAM, and 'PM951 NVMe SAMSUNG 256GB' NVMe-SSD, in case if it matters. I am going to create another filesystem on this pool, set it as bootfs in zpool property, and install a pristine OS there to see what happens. In the meantime, am I safe to run: # freebsd-upgrade rollback while the system is up in kernel.old (11.1p10) mode, to get my system back to boot 11.1p10 by default? Thank you in advance. Hiroharu