0

Until yesterday, I have run Linux kernel 4.9 with zfs / spl 0.6.9.5 (I believe). zpool status and other diagnostic commands did not show any error with the ZFS pool (I am only having one pool).

Today, I have upgraded the kernel to 4.16 and zfs / spl to 0.7.8. But I can't boot the new kernel because zfs claims that the pool is damaged during the boot procedure, leaving me at the busybox command prompt (I have the root file system on ZFS).

The detailed error number is zfs-8000-5e and is described in detail here. According to that description, all data has been lost (well, not really, because I have backups of course, but this would have been a real pain).

Fortunately, I have made my experiences already and am always keeping my old kernel versions and modules. So I was able to boot the previous kernel, and guess what: The pool has been imported during boot without any problem as usual. So the pool is definitely not damaged or corrupt.

Now I have two questions:

1) Why does ZFS (with the new kernel) claim that I have lost all my data although that pool is working perfectly?

2) How can I use that pool with the new kernel (without restoring from backup - I will definitely not destroy my data and recover from backup just to be able to use a newer zfs version)?

  • Have you updated you zfs-initramfs as well? Which linux distribution are you using? – Michael D. Apr 30 '18 at 9:06
  • @Michael D Yes, of course. The modules get loaded correctly when booting the new kernel (lsmod is provided in the busybox I hit). I am using debian stretch. – Binarus Apr 30 '18 at 9:28
  • Have you checked the bug reports, yet? bugs.debian.org/cgi-bin/pkgreport.cgi?src=zfs-linux – Michael D. Apr 30 '18 at 14:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.