Pavel Zhukov has posted comments on this change.
Change subject: Add additional vgcheck to fix vgextend workflow on HSMs.
......................................................................
Patch Set 9:
I don't understand how this fixes the related bug. If the issue was a missing pv, how running vgchk again fixes the missing pv?
The issue is stale filters
If the issue is stale filters, why not invalidate the filters in _reloadvgs, when we find that a vg has missing pv? Why wait until selftest()?
Because _reloadvgs uses vgs (not vgck). VGs returns zero even if PV is missed (see commit message) and We know nothing about missed PVs on this step. I thought about rebuilding filters before vgs command, but do we really want to do it each time for each vg?
--
To view, visit
http://gerrit.ovirt.org/20552
To unsubscribe, visit
http://gerrit.ovirt.org/settings
Gerrit-MessageType: comment
Gerrit-Change-Id: If1eeed1c203f2c8c73370987048565d665932299
Gerrit-PatchSet: 9
Gerrit-Project: vdsm
Gerrit-Branch: master
Gerrit-Owner: Pavel Zhukov
pzhukov@redhat.com
Gerrit-Reviewer: Dan Kenigsberg
danken@redhat.com
Gerrit-Reviewer: Eduardo
ewarszaw@redhat.com
Gerrit-Reviewer: Federico Simoncelli
fsimonce@redhat.com
Gerrit-Reviewer: Lee Yarwood
lyarwood@redhat.com
Gerrit-Reviewer: Nir Soffer
nsoffer@redhat.com
Gerrit-Reviewer: Pavel Zhukov
pzhukov@redhat.com
Gerrit-Reviewer: Sergey Gotliv
sgotliv@redhat.com
Gerrit-Reviewer: Tomáš Došek
tdosek@redhat.com
Gerrit-Reviewer: Yeela Kaplan
ykaplan@redhat.com
Gerrit-Reviewer: oVirt Jenkins CI Server
Gerrit-HasComments: No