[Bug 232671] [gmirror] gmirror fails to recover from degraded mirror sets in some circumstances
bugzilla-noreply at freebsd.org
bugzilla-noreply at freebsd.org
Tue Nov 27 05:40:16 UTC 2018
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=232671
Conrad Meyer <cem at freebsd.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|New |In Progress
Assignee|geom at FreeBSD.org |cem at freebsd.org
URL| |https://reviews.freebsd.org
| |/D18062
--- Comment #7 from Conrad Meyer <cem at freebsd.org> ---
This can be boiled down to the following problem:
1. sc_ndisks is 2
2. we have three eligible components:
i. da2p5 (the stale mirror that was ejected)
ii. da16p3 (the mirror that is partially synchronized)
iii. da15p3 (the only good / "ACTIVE" mirror in the set)
3. STARTING auto-transitions to running based on sc_ndisks == len(sc->sc_disks)
*prior* to removing stale components.
I think the working patch in https://reviews.freebsd.org/D18062 will eliminate
(3), rejecting (i) when (ii) is added, preventing auto-start until (iii) is
tasted, or the timeout expires.
--
You are receiving this mail because:
You are the assignee for the bug.
More information about the freebsd-geom
mailing list