[OpenAFS-devel] release-team meeting minutes 2014-02-19

Stephan Wiesand stephan.wiesand@desy.de
Fri, 21 Feb 2014 18:35:00 +0100


Log: =
http://conference.openafs.org/release-team@conference.openafs.org/2014-02-=
19.txt

Participants:

* Andrew Deason
* Ben Kaduk
* Derrick Brashear
* Jeffrey Altman
* Marc Dionne
* Mike Meffie
* Stephan Wiesand

=3D=3D Problem reports =3D=3D

=
https://lists.openafs.org/pipermail/openafs-info/2014-February/040506.html=


Fssync problems during vos backupsys on NBSD. Andrew is working with the =
reporter. Not a  blocker for the next release.

=
https://lists.openafs.org/pipermail/openafs-info/2014-February/040503.html=


Linux panics due to multiple mounts. Not much progress. Old problem, no =
solution at hand yet. Not a blocker for the time being.

RT #131804 (vos release -stayonline)

New feature in 1.6.2. There's another report that it doesn't work. =
Broken by a later patch to vos?

RT #131578 (afs commander crashes, reported a year ago)

Needs work, may hinder AFS adoption on OS X. Not a blocker.

=3D=3D Linux news =3D=3D

Marc tested with the current mainline this morning, no need for changes =
yet.

=3D=3D 1.6.7^H8 release =3D=3D

1.6.7 will be a security release (1.6.6 plus a security fix, nothing =
else). Hence what was planned to be released as 1.6.7 will become 1.6.8.

We plan for a 1.6.8pre1 shortly before EAKC. Thus, further changes to be =
considered for 1.6.8 should be either very simple or urgent. Complex =
non-critical fixes or additional features should be postponed to 1.6.9. =
A possible fix for the linux multiple mounts issue
will still be considered for inclusion, as well as gerrit 9020 (vos =
release -force-reclone), if ready in time.

More review is needed for many changes lined up for inclusion in gerrit. =
Please focus on those not marked -2 for the time being.

=3D=3D 1.6.8 changes pre announcements =3D=3D

Jeffrey to take care. The ih_sync change to default to onclose will be =
included.

=3D=3D Status of items foreseen to be included in 1.6.8 =3D=3D

The "give up callback" changes were recently rebased but no longer =
build. Stephan will try to fix them. [update: they now build, but what =
we have seems not to work as expected]


[the remainder of this text is 1.6.8 bookeeping]

Idledead change to help the "no buffer space available" problem: Gerrit =
10773 ist still being reviewed on master, stalled on two trivial changes =
to be done.

Parallel build fixes =
(http://gerrit.openafs.org/#q,project:openafs+branch:master+topic:parallel=
-make,n,z ). Stalled.

=3D=3D=3D Accepted on master, to be pulled up to 1_6_x =3D=3D=3D

- nothing left that was discussed before -

=3D=3D=3D Pulled up to 1_6_x, to be reviewed =3D=3D=3D

Ben's changes to allow using the OS X buildbot slave for 1_6_x too:=20
10744 afs: Handle reading past the end of a file (new patch set)

Interrupt clients accessing offlining volumes:
10799 viced: Add options for interrupting clients
6266  vol: Interrupt RX calls accessing offlining vols (single +1 yet)

Give up callbacks at shutdown:
8840  libafs: Correct GUACB afs_LoopServers flags
6272  unix: giveupallcallbacks at shutdown

Others (no progress, or insufficient progress to merge):
10772 vol: Nuke parent vol special inodes
10771 salvager: Ignore linktable-only RW volumes
10770 namei: Set inconsistent linktable linkCount to 0
10759 viced: Avoid issuing redundant TMAY requests
10757 viced: Improve client error log messages
10755 viced: Avoid endless BCB loop

--=20
Stephan Wiesand
DESY -DV-
Platanenenallee 6
15738 Zeuthen, Germany