Fedora 18 Final blocker status: required fixes, karma, etc

[Adam Williamson]

Time for the first Fedora 18 Final blocker status update. This will be a
big one, folks, strap yourselves in! As usual, this is an attempt to let
everyone know where we are with the current blocker list: what bugs we
still need fixes for, where we need further testing, karma and/or
information. 

To hit the current schedule we need to basically be done - have a viable
RC - by the end of this week. *All* the accepted blockers at least need
to be addressed by then.

Here is a summary of required action, for those who don't have time to
read the whole thing:

Summary
-------

The following updates need testing and karma:

* https://admin.fedoraproject.org/updates/anaconda-18.37.3-1.fc18
* https://admin.fedoraproject.org/updates/shim-signed-0.2-3.2.fc18
* https://admin.fedoraproject.org/updates/lorax-18.24-1.fc18
* https://admin.fedoraproject.org/updates/cloud-init-0.7.1-2.fc18
* https://admin.fedoraproject.org/updates/device-mapper-multipath-0.4.9-36.fc18
* https://admin.fedoraproject.org/updates/FEDORA-2012-20368/plymouth-0.8.8-5.fc18

The following bugs require developer attention:

anaconda team
-------------

* https://bugzilla.redhat.com/show_bug.cgi?id=858628
* https://bugzilla.redhat.com/show_bug.cgi?id=873224
* https://bugzilla.redhat.com/show_bug.cgi?id=875944
* https://bugzilla.redhat.com/show_bug.cgi?id=876716
* https://bugzilla.redhat.com/show_bug.cgi?id=879187
* https://bugzilla.redhat.com/show_bug.cgi?id=885378
* https://bugzilla.redhat.com/show_bug.cgi?id=885912
* https://bugzilla.redhat.com/show_bug.cgi?id=887236
* https://bugzilla.redhat.com/show_bug.cgi?id=877658 (proposed)
* https://bugzilla.redhat.com/show_bug.cgi?id=883699 (proposed)
* https://bugzilla.redhat.com/show_bug.cgi?id=886647 (proposed)
* https://bugzilla.redhat.com/show_bug.cgi?id=887539 (proposed)
* https://bugzilla.redhat.com/show_bug.cgi?id=888089 (proposed)
* https://bugzilla.redhat.com/show_bug.cgi?id=886061 (proposed)

haraldh
-------

* https://bugzilla.redhat.com/show_bug.cgi?id=874486
* https://bugzilla.redhat.com/show_bug.cgi?id=860525
* https://bugzilla.redhat.com/show_bug.cgi?id=881670
* https://bugzilla.redhat.com/show_bug.cgi?id=880271
* https://bugzilla.redhat.com/show_bug.cgi?id=881670

others
------

* https://bugzilla.redhat.com/show_bug.cgi?id=830434 - NetworkManager folks
* https://bugzilla.redhat.com/show_bug.cgi?id=876218 - systemd folks
* https://bugzilla.redhat.com/show_bug.cgi?id=882212 - systemd folks
* https://bugzilla.redhat.com/show_bug.cgi?id=873817 - jciesla
* https://bugzilla.redhat.com/show_bug.cgi?id=875846 - lorax folks

The following bugs require further information or testing from QA
folks / reporters aside from karma:

* https://bugzilla.redhat.com/show_bug.cgi?id=880263 - Mark Hamzy
* https://bugzilla.redhat.com/show_bug.cgi?id=876789 - Jes Sorensen
* https://bugzilla.redhat.com/show_bug.cgi?id=886314 - boblfoot or others
* https://bugzilla.redhat.com/show_bug.cgi?id=882212 - Jan Vcelak

Accepted blockers
-----------------

1. https://bugzilla.redhat.com/show_bug.cgi?id=830434 - "Omits
"IPV6INIT=yes" from kickstart-generated ifcfg-* file"

This needs fixing in NetworkManager's ifcfg-rh plugin writer. jirka has
posted a patch in the bug, it needs applying and we need a
NetworkManager update that includes it.

2. https://bugzilla.redhat.com/show_bug.cgi?id=858628 - "Some buttons
and labels in Anaconda can't be localized"

This became a kind of catch-all bug. The last bit that Chris really
figures is a part of this bug now has a patch posted, so should wind up
in an anaconda build soon, then we can close it out.

3. https://bugzilla.redhat.com/show_bug.cgi?id=873224 - "TypeError:
cannot concatenate 'str' and 'NoneType' objects"

dlehman is still working on this one, as of 4 days ago. We need him to
finish working on it and push it to an anaconda build.

4. https://bugzilla.redhat.com/show_bug.cgi?id=874486 - "progress
indicator for mediacheck isn't displayed, so users may think the
installer is hung"

Harald has posted a patch for this, but we are still waiting on a
build/update including the patch so far as I know.

5. https://bugzilla.redhat.com/show_bug.cgi?id=875944 - "Shrink function
in 'Reclaim space' dialog does not allow user to specify target size,
automatically picks smallest possible size"

We need anaconda team to decide what to do about this, and do it.

6. https://bugzilla.redhat.com/show_bug.cgi?id=876218 - "kernel boot +
nfsiso repo = hang on reboot"

We're still trying to track this down, really: it needs some input from
the systemd developers. Kay, Lennart?

7. https://bugzilla.redhat.com/show_bug.cgi?id=876716 - "anaconda
crashes after setting root password"

This is a very tricky threading issue between anaconda and yum and RPM.
It's under discussion on the anaconda list ATM, but we need to pick a
fix and go with it.

8. https://bugzilla.redhat.com/show_bug.cgi?id=879187 - "NFS in RC1
regressed vs. TC9: does not override inst.stage2 when passed as
inst.repo, interactive NFS is impossible"

sbueno is supposed to be working on this one, we are waiting on the fix.

9. https://bugzilla.redhat.com/show_bug.cgi?id=880263 - "DMError:
partition activation failed for 'mpatha'"

This should be fixed in TC3, needs to be verified and the anaconda
update needs karma. Mark, can you check this for us?

10. https://bugzilla.redhat.com/show_bug.cgi?id=885378 - "Install to
existing encrypted LV fails (AttributeError: 'NoneType' object has no
attribute 'name')"

This one has been a blocker for a week without any movement. We need
anaconda team to take a look at it.

11. https://bugzilla.redhat.com/show_bug.cgi?id=885912 - "Resize of NTFS
partition results in partition smaller than the filesystem, broken
Windows install"

clumens is looking into this one, we need a fix to test.

12. https://bugzilla.redhat.com/show_bug.cgi?id=887236 -
"UnicodeDecodeError: 'ascii' codec can't decode byte 0xe6 in position 3:
ordinal not in range(128)"

We need anaconda team to look into this one. Probably python problems.

13. https://bugzilla.redhat.com/show_bug.cgi?id=887363 - "cloud-init
0.7.1 change prevents creation of ec2-user"

This is to do with EC2 usage. A fix is available, we need EC2 users to
verify that it works and provide karma on
https://admin.fedoraproject.org/updates/FEDORA-2012-20389/cloud-init-0.7.1-2.fc18 .

14. https://bugzilla.redhat.com/show_bug.cgi?id=887357 - "workaround for
for appliance creation in koji"

This is to do with creating appliance images, particularly the EC2
images we create these days. Need releng to confirm the fix and provide
karma here, I think. fix is
https://admin.fedoraproject.org/updates/device-mapper-multipath-0.4.9-36.fc18

15. https://bugzilla.redhat.com/show_bug.cgi?id=880271 - "Wrong keyboard
layout in initramdisk"

This one has also been blocker for a week with no movement, needs dracut
team to look into it.

16. https://bugzilla.redhat.com/show_bug.cgi?id=883075 - "fedup
upgrading is too quiet"

There is actually a proposed fix now:
https://admin.fedoraproject.org/updates/FEDORA-2012-20368/plymouth-0.8.8-5.fc18 . Will suggests a test method in https://bugzilla.redhat.com/show_bug.cgi?id=879295#c31 . We need to test and karma the update.

17. https://bugzilla.redhat.com/show_bug.cgi?id=873817 - "inkscape
packaging bugs prevent install"

We thought this was fixed, but it appears to still happen in TC3. We
need the maintainer (Jon Ciesla) to see if they can figure out why it's
still broken.

18. https://bugzilla.redhat.com/show_bug.cgi?id=876789 -
"FormatDestroyError: error wiping old signatures
from /dev/md/Volume0_0p2: 1"

The fix for this should be in TC3, so it needs re-testing. As Jes is
best able to reproduce this, it'd help if he could test it. Fix is
https://admin.fedoraproject.org/updates/FEDORA-2012-19873/lvm2-2.02.98-4.fc18 and is in stable - if the fix is confirmed the bug could be closed.

Proposed blockers
-----------------

1. https://bugzilla.redhat.com/show_bug.cgi?id=877658 - "[i18n] some
storage-related error messages not marked for translation: "Not enough
free space on disks for automatic partitioning""

This is a more specific translation issue. Always hard to know where to
draw the line on these, but it's likely to be accepted as at least NTH,
so it'd be good if anaconda team can look at it.

2. https://bugzilla.redhat.com/show_bug.cgi?id=883699 - "Changing LV
name in custom partitioning crashes installer (TypeError: Argument 2
does not allow None as a value)"

Seems to be related to changing the name of an LV in custom
partitioning. Likely to be accepted as at least NTH, so be good if
anaconda team can work this into an update (a fix has been posted).

3. https://bugzilla.redhat.com/show_bug.cgi?id=886314 - "MDRaidError:
md_node_from_name failed: [Errno 2] No such file or directory:
'/dev/md/root'"

Seems related to certain software RAID-10 installation scenarios, but
dlehman could not reproduce. We need to figure out more details on the
cause.

4. https://bugzilla.redhat.com/show_bug.cgi?id=886647 - "Anaconda throws
exception trying to setup network"

It seems like we need more info on the cause here to determine blocker
status, but the reporter has provided the info requested by vpodzime, so
it may be up to anaconda team to determine cause at this point.

5. https://bugzilla.redhat.com/show_bug.cgi?id=887201 - "unable to place
the root filesystem(/) on a xfs(ext3)formatted partition"

This one is tested fixed in current anaconda build, so it just needs
karma and can be closed out. Would probably be accepted as a blocker.

6. https://bugzilla.redhat.com/show_bug.cgi?id=887539 -
"FormatDestroyError: error wiping old signatures from /dev/sdb2: 1"

We are still trying to determine the precise situation here, and whether
it's a dupe of an existing bug or a different case. Need input from the
anaconda team.

7. https://bugzilla.redhat.com/show_bug.cgi?id=888089 - "ValueError: A
RAID0 set requires at least 2 members"

This seems to be fairly clearly explained, and I'd guess it'd be
accepted as a blocker. Need anaconda team to look at fixing it.

8. https://bugzilla.redhat.com/show_bug.cgi?id=860525 -
"dracut-pre-pivot[272]: mount: wrong fs type, bad option, bad superblock
on /var/lib/nfs/rpc_pipefs,"

This one still isn't really clear at all, but seems to be somehow to do
with mount options on NFS installs. It would help if Harald could
clarify exactly what's going on here - he seems to understand, but
hasn't explained.

9. https://bugzilla.redhat.com/show_bug.cgi?id=886061 - "core fedora
repository error must be fatal, otherwise the upgraded system might be
broken heavily"

This is a potential problem with fedup upgrades if contacting the main
repo somehow fails. We need input from wwoods on how bad the impact of
this may be in the shipped configuration (see
https://bugzilla.redhat.com/show_bug.cgi?id=886061#c10 )

10. https://bugzilla.redhat.com/show_bug.cgi?id=875846 - "[i18n] "ON"
and "OFF" not translated in switches on DVD (gtk30.mo not on DVD)"

Another translation issue. This one may be obvious enough that it
doesn't make blocker, but I'm guessing there. Issue seems to be fairly
clearcut and is likely to make NTH at least, so it would be good if the
lorax maintainer can fix it up.

11. https://bugzilla.redhat.com/show_bug.cgi?id=888026 - "shim not found
when doing composing"

This is in fact fixed already with a new shim-signed build, but the
build needs karma. Just the fact that TC3 is composed proves this bug is
fixed: that's probably enough to up-karma the build, as even if SB
support was somehow completely broken, that wouldn't be a _regression_,
so we may as well go ahead and push the shim-signed build to stable. So
QA/releng folks should up-karma the update.

12. https://bugzilla.redhat.com/show_bug.cgi?id=881670 - "Provide fedup
hook to add x-systemd.device-timeout=0 to mount options for encrypted
file systems that need a passphrase from the user"

We're still waiting for some input from Harald here to evaluate blocker
status. Hard to make a guess here.

13. https://bugzilla.redhat.com/show_bug.cgi?id=882212 - "localectl
set-x11-keymap: variant settings does not work"

This affects keyboard setting migration on upgrade apparently. Not clear
whether this is yum only or also affects fedup, but even if it's
yum-only, probably will be accepted as NTH, so it would be good to get a
systemd build with this fix, systemd folks.

14. https://bugzilla.redhat.com/show_bug.cgi?id=861280 - "KDE Crashes or
Hangs After Moving Konqueror Window Quickly"

This Bug Is Obviously Very Significant As It Has Been Capitalized. Looks
like a single-system/single-adapter X bug, which by policy we don't take
as blockers (or else we'd never release anything). I'd guess this will
be rejected.

Thanks for your attention to these bugs everyone! Thanks.
-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
http://www.happyassassin.net