[Bug 207667] Reproducable kernel panics whilst scrubbing a zvol pool...

bugzilla-noreply at freebsd.org bugzilla-noreply at freebsd.org
Wed Apr 6 17:10:10 UTC 2016


--- Comment #9 from Andriy Gapon <avg at FreeBSD.org> ---
(In reply to Michelle Sullivan)

Replying to multiple comments at once.

Yes, taking a picture of a monitor is a fine last measure.  A serial console,
real or emulated via, e.g., IPMI or AMT, is better, of course.

The panic looks like a result of a well-known (among developers) and very hard
to fix problem with interaction between GEOM and ZVOLs that arises from using
ZFS pools on top of ZFS volumes.  I can not offer you any fix although I have
got some work-in-progress on this issue.

I can suggest a work-around: create a FreeBSD VM using the volumes comprising
the pool as the VM's disks and run the scrub within the VM.

Besides zpool scrub commands like zpool reopen, zpool split, zpool online [-e]
can also trigger the same bug for pools on top of ZFS volumes.

You are receiving this mail because:
You are on the CC list for the bug.

More information about the freebsd-amd64 mailing list