Date: Thu, 25 Oct 2018 01:08:16 +0000 From: bugzilla-noreply@freebsd.org To: geom@FreeBSD.org Subject: [Bug 232671] [gmirror] gmirror fails to recover from degraded mirror sets in some circumstances Message-ID: <bug-232671-14739-pGnhr7pgiR@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-232671-14739@https.bugs.freebsd.org/bugzilla/> References: <bug-232671-14739@https.bugs.freebsd.org/bugzilla/>
next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D232671 --- Comment #1 from Conrad Meyer <cem@freebsd.org> --- My proposed fix is, in g_mirror_update_device: don't transition from STARTI= NG to RUNNING unless we've got at least one ACTIVE mirror. What is a gmirror device going to do with zero active mirrors? It seems useless. Possible mitigations for similar classes of issue: * When a stale mirror is detected at runtime *and* we've already got a comp= lete mirrorset (#6 above), zero out the gmirror superblock or otherwise disassoc= iate it from the logical mirrorset. It's not going to get less stale if we igno= re it. * In g_mirror_update_device when we're in RUNNING and notice we have no mirrors, transition back to STARTING? This sounds a little silly to me; we should probably just avoid transitioning to RUNNING unless we have a valid mirror. --=20 You are receiving this mail because: You are the assignee for the bug.=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-232671-14739-pGnhr7pgiR>