Fedora 8 blocker update - testers needed (still!)

[Will Woods]

As promised, here's an update on the current Fedora 8 Blocker list. 

As a note: the deadline for blocker fixes is Tuesday, October 29. If we
still have unfixed critical bugs Tuesday, the release will have to be
delayed.

But I don't think that'll happen. Things are looking OK. We currently
have 14 bugs on the blocker list, which I'll talk about in detail below.
About half of them have fixes that are waiting for confirmation, and a
couple of them could be dropped if necessary because they only affect
particular, uncommon hardware.

What we need help testing is:

1) ALSA - 1.0.15 was released last week, and we decided it was a
low-risk update that would fix audio problems. for a lot of people. You
probably saw mail from Chuck Ebbert about this already.

2) Intel wireless - John Linville, our fearless (and handsome) wireless
maintainer, just added a fix for iwlwifi that should help some Intel
wireless users. 

Both of the above patches are in kernel-2.6.23.1-37.fc8, which should
land in rawhide tomorrow. You can get it from koji in the meantime:
http://koji.fedoraproject.org/koji/buildinfo?buildID=22511

3) dmraid, especially Intel
See bug 349161 below. Something is causing mkinitrd to mess up the /init
script, which leaves the reporter's system unbootable. We can't
reproduce it at all. 

If you've got dmraid (aka BIOS RAID aka fakeraid) hardware, ESPECIALLY
the one mentioned in this bug (ICH9R chipset + Intel Matrix Raid),
please try to install rawhide and let us know if it works or not.

Here's the full blocker list, in detail:

=== Miscellaneous bugs from last time ===

* Bug 208607: ASSIGNED - kernel maintainers
- mkinitrd: iSCSI root requires crc32c module

I believe we plan to fix this by building crc32c into the kernel for
release (which would make this a non-blocker). We'll hopefully have a
smarter fix for F9 that will actually close the bug.


* Bug 242766: ASSIGNED - alan
- [pata_via via82c596a] hard drive not found during install attempt

All testing with pata_via seems to indicate that it works for most
people. Alan's debugging efforts indicate that this is a drive-specific
failure on a machine with a buggy BIOS. Fun. This will be dropped from
the blocker list, since it works under normal circumstances.

(By the way- thanks to everyone who answered the call for testing here.
You guys rock!)


* Bug 335161: ASSIGNED - triad (Linus Walleij)
- multilib file conflicts in gnokii

Fix was incomplete, needs a rebuild. This prevents people from being
able to yum upgrade from F8Test3 to F8 final. 


* Bug 330931: MODIFIED - katzj
- liveinst fails - sg devices confuse anaconda?

This was waiting on some info from the reporter, which has been
provided. It involves some uncommon hardware so this bug may be dropped
from the blocker list if it can't be fixed easily.


=== NetworkManager bugs ===

* Bug 311511: NEW - dcbw
- x86_64 iwl3945 NM crash after Keyring loop Deny

Comments indicate that the crash is fixed, but it still loops on bad
passwords. 


* Bug 323371: NEEDINFO - dcbw
- NM 0.7 doesn't support WPA/WPA2 Enterprise (certificate-based
authentication)

Reporter seems to indicate that this is mostly fixed, but some testing
on a fresh install is needed to confirm the fix.


* Bug 353091: MODIFIED - dcbw
- NetworkManager-0.7.0-0.3.svn3020.fc8 crashes when trying to switch to
wireless Network

nm-applet crasher. A fix will be in the next build, and then this can be
retested and (hopefully) closed. 


=== Added since Wednesday ===

* Bug 349161: NEW - pjones
- Missing entries in the init script of the initrd image makes the
system unbootable on dmraid

This is a mysterious one. mkinitrd is creating a bad /init script for a
dmraid (aka fakeraid, aka BIOS RAID) raid0 set. But we can't reproduce
it on any of the dmraid machines we've tried.

It's a fixable problem, and I don't think the hardware is very common,
and we can't reproduce the bug ourselves, so.. if push comes to shove
this bug might get dropped from the blocker list and fixed post-F8.


* Bug 352511: MODIFIED - dwalsh
- return of an avc on useradd with livecd user

Should be fixed in selinux-policy -33 - we'll confirm as soon as Jeremy
makes some Live images.


* Bug 353571: MODIFIED - dwalsh
- /usr/bin/crontab blocked by selinux


Apparently fixed in selinux-policy-3.0.8-36; waiting for confirmation
from the reporter.

* Bug 353961: NEW - twoerner
- lokkit doesn't write iptables config with --nostart

This was reported earlier this morning - no response from the maintainer
yet. Technically I don't think this breaks any of the ReleaseCriteria,
so it could be dropped. But proper firewall function is important for
security so we want to keep an eye on this.


* Bug 354531: ASSIGNED - dwmw2
- script failure

This is a simple scriptlet problem with bluez-utils, but it keeps the
'bluetooth' service from being installed properly. Improved Bluetooth
support is one of the major features of F8, so we want this to work
right. The maintainer is aware of the problem and working to fix it.


* Bug 354571: MODIFIED - anaconda maintainers
- live install fails when there's a pre-existing install

Fixed in anaconda CVS - the next set of Live images should allow us to
retest and close this.


* Bug 354781: NEW - skvidal
- yum-skip-broken changes yum's dependency solver even when it's not
active

This is a bug with the popular yum-skip-broken plugin that could cause
it to make upgrades from F7 fail. Upgrade failures are bad. 



And there you have it. Most of those bugs already have fixes pending,
and none of the rest are hair-on-fire critical bugs, so I think we're in
good shape. If you can help test the stuff I mentioned above, or you
have found some huge problem that we've missed, please let us know.

Thanks for helping make Fedora more awesome!

-w

[Mike Chambers]

On Fri, 2007-10-26 at 19:18 -0400, Will Woods wrote:
>  As a note: the deadline for blocker fixes is Tuesday, October 29. If
>  we
>  still have unfixed critical bugs Tuesday, the release will have to be
>  delayed. 
Tuesday is Oct 30th.  So is it Mon the 29th, or Tue the 30th? hehe

-- 
Mike Chambers
Madisonville, KY

"Best lil town on Earth!"

[Will Woods]

On Fri, 2007-10-26 at 19:16 -0500, Mike Chambers wrote:
>  On Fri, 2007-10-26 at 19:18 -0400, Will Woods wrote:
>  > As a note: the deadline for blocker fixes is Tuesday, October 29. If
>  > we
>  > still have unfixed critical bugs Tuesday, the release will have to be
>  > delayed.
>  
>  Tuesday is Oct 30th.  So is it Mon the 29th, or Tue the 30th? hehe 
Heh sorry. It's Tuesday, the 30th. 

(I was originally just going to lie and say it was Monday, because I
*know* someone will want to get Just One Last Fix in, but I changed my
mind. Unfortunately I didn't change the date.)

-w

[Neal Becker]

Tested kernel-2.6.23.1-26.fc8.x86_64

Sound working great! (IHC8)
00:1b.0 Audio device: Intel Corporation 82801H (ICH8 Family) HD Audio Controller (rev 03)

iwl4965 working great!
02:00.0 Network controller: Intel Corporation PRO/Wireless 4965 AG or AGN Network
Connection (rev 61)

Thanks - you guys rock!

[Christopher Aillon]

Neal Becker wrote:
>  Tested kernel-2.6.23.1-26.fc8.x86_64 
Which is much older than the kernel we want tested.  It would be good to 
know that the newer kernels work, too.  :-)


>  Sound working great! (IHC8)
>  00:1b.0 Audio device: Intel Corporation 82801H (ICH8 Family) HD Audio Controller (rev
> 03)
>  
>  iwl4965 working great!
>  02:00.0 Network controller: Intel Corporation PRO/Wireless 4965 AG or AGN Network
> Connection (rev 61)
>  
>  Thanks - you guys rock!
>

[Neal Becker]

Christopher Aillon wrote:

>  Neal Becker wrote:
> > Tested kernel-2.6.23.1-26.fc8.x86_64
>  
>  Which is much older than the kernel we want tested.  It would be good to
>  know that the newer kernels work, too.  :-)
>  
>  
> > Sound working great! (IHC8)
> > 00:1b.0 Audio device: Intel Corporation 82801H (ICH8 Family) HD Audio
> > Controller (rev 03)
> > 
> > iwl4965 working great!
> > 02:00.0 Network controller: Intel Corporation PRO/Wireless 4965 AG or AGN
> > Network Connection (rev 61)
> > 
> > Thanks - you guys rock!
> > 
>   
I'm sorry That should be:
tested: kernel-2.6.23.1-37.fc8.x86_64

[Alan Cox]

On Fri, Oct 26, 2007 at 07:18:58PM -0400, Will Woods wrote:
>  3) dmraid, especially Intel
>  See bug 349161 below. Something is causing mkinitrd to mess up the /init
>  script, which leaves the reporter's system unbootable. We can't
>  reproduce it at all.  
Is this the same as the open FC7 reports about initrd problems with dmraid
and LVM etc that are sitting there one with a patch that claims to fix it ?

Alan

[Will Woods]

On Fri, 2007-10-26 at 20:50 -0400, Alan Cox wrote:
>  On Fri, Oct 26, 2007 at 07:18:58PM -0400, Will Woods wrote:
>  > 3) dmraid, especially Intel
>  > See bug 349161 below. Something is causing mkinitrd to mess up the /init
>  > script, which leaves the reporter's system unbootable. We can't
>  > reproduce it at all. 
>  
>  Is this the same as the open FC7 reports about initrd problems with dmraid
>  and LVM etc that are sitting there one with a patch that claims to fix it ? 
Probably not? Most of the F7 bugs with initrd & dmrait were related to
the changeover to libata and the initrd being wrong because of device
names changing and such. 

But if you've got some specific bug IDs, I'd definitely like to take
look at 'em..

-w

[Michael Wiktowy]

On 10/26/07, Will Woods <wwoods(a)redhat.com&gt; wrote:
>  On Fri, 2007-10-26 at 20:50 -0400, Alan Cox wrote:
>  > On Fri, Oct 26, 2007 at 07:18:58PM -0400, Will Woods wrote:
>  > > 3) dmraid, especially Intel
>  > > See bug 349161 below. Something is causing mkinitrd to mess up the /init
>  > > script, which leaves the reporter's system unbootable. We can't
>  > > reproduce it at all.
>  >
>  > Is this the same as the open FC7 reports about initrd problems with dmraid
>  > and LVM etc that are sitting there one with a patch that claims to fix it ?
>
>  Probably not? Most of the F7 bugs with initrd & dmrait were related to
>  the changeover to libata and the initrd being wrong because of device
>  names changing and such.
>
>  But if you've got some specific bug IDs, I'd definitely like to take
>  look at 'em.. 
I know that I reported some dmraid+LVM flakiness back in F7 test cycle
that wasn't resolved then.

https://bugzilla.redhat.com/show_bug.cgi?id=234719
and
https://bugzilla.redhat.com/show_bug.cgi?id=234724

Unfortunately, I finally made the decision to proceed with a stable
installation and abandon using dmraid and don't have spare drive space
to test this with any more.

I'll see if I can scrounge up some drives to do some testing again ...
but it seems that dmraid is an issue that gets attention in the dev
cycle right at the point where it is too late to do anything about it
because folks hope it will just go away.

/Mike

[Alan Cox]

On Fri, Oct 26, 2007 at 11:40:56PM -0400, Michael Wiktowy wrote:
>  > Probably not? Most of the F7 bugs with initrd & dmrait were
> related to
>  > the changeover to libata and the initrd being wrong because of device
>  > names changing and such.
>  >
>  > But if you've got some specific bug IDs, I'd definitely like to take
>  > look at 'em.. 
This broke mid F7.

241949 244535 237415 are the ones I have scribbled down

Not sure precisely which as buggyzilla is down

[John Summerfield]

Will Woods wrote:
>  As promised, here's an update on the current Fedora 8 Blocker
> list. 
>  
>  As a note: the deadline for blocker fixes is Tuesday, October 29. If we
>  still have unfixed critical bugs Tuesday, the release will have to be
>  delayed. 
Did I miss something, or do you not intend to fix 
https://bugzilla.redhat.com/show_bug.cgi?id=251493

It seems it also applies to ESPRIMO 5720, see the "Fedora 8 test 3 
Hangs" thread.




-- 

Cheers
John

-- spambait
1aaaaaaa(a)coco.merseine.nu  Z1aaaaaaa(a)coco.merseine.nu
-- Advice
http://webfoot.com/advice/email.top.php
http://www.catb.org/~esr/faqs/smart-questions.html
http://support.microsoft.com/kb/555375

Please do not reply off-list

[Chuck Anderson]

On Fri, Oct 26, 2007 at 07:18:58PM -0400, Will Woods wrote:
>  * Bug 323371: NEEDINFO - dcbw
>  - NM 0.7 doesn't support WPA/WPA2 Enterprise (certificate-based
>  authentication)
>  
>  Reporter seems to indicate that this is mostly fixed, but some testing
>  on a fresh install is needed to confirm the fix. 
I updated this bug since you composed this message.  The new WPA 
Enterprise support isn't yet working for me with PEM certificates.  I 
think it is important to get this case working, as many sites use 
cert-based authentication for wireless, and it used to work in earlier 
releases.

[Bill Nottingham]

>  * Bug 330931: MODIFIED - katzj
>  - liveinst fails - sg devices confuse anaconda?
>  
>  This was waiting on some info from the reporter, which has been
>  provided. It involves some uncommon hardware so this bug may be dropped
>  from the blocker list if it can't be fixed easily. 
Should be fixed in currently built anaconda.

>  * Bug 353961: NEW - twoerner
>  - lokkit doesn't write iptables config with --nostart
>  
>  This was reported earlier this morning - no response from the maintainer
>  yet. Technically I don't think this breaks any of the ReleaseCriteria,
>  so it could be dropped. But proper firewall function is important for
>  security so we want to keep an eye on this. 
Should already be fixed and tagged in 1.0.8-3.fc8.

Bill

[Seth Vidal]

On Fri, 2007-10-26 at 19:18 -0400, Will Woods wrote:

>  * Bug 354781: NEW - skvidal
>  - yum-skip-broken changes yum's dependency solver even when it's not
>  active
>  
>  This is a bug with the popular yum-skip-broken plugin that could cause
>  it to make upgrades from F7 fail. Upgrade failures are bad.  
Neither I nor the reporter (notting) can make this happen again and
looking at the code which runs in skip-broken it isn't even possible so
something else happened during that run where he noticed the issue but
it's not this bug. So this one is closed.

-sv

[Matthew Saltzman]

On Fri, 2007-10-26 at 19:18 -0400, Will Woods wrote:
>  As promised, here's an update on the current Fedora 8 Blocker
> list. 
>  
>  As a note: the deadline for blocker fixes is Tuesday, October 29. If we
>  still have unfixed critical bugs Tuesday, the release will have to be
>  delayed.
>  
>  But I don't think that'll happen. Things are looking OK. We currently
>  have 14 bugs on the blocker list, which I'll talk about in detail below.
>  About half of them have fixes that are waiting for confirmation, and a
>  couple of them could be dropped if necessary because they only affect
>  particular, uncommon hardware.
>  
>  What we need help testing is:
>  
>  1) ALSA - 1.0.15 was released last week, and we decided it was a
>  low-risk update that would fix audio problems. for a lot of people. You
>  probably saw mail from Chuck Ebbert about this already. 
kernel-2.6.23.1-37.fc8: 

Most sound works well on my ThinkPad T61, except that the volume buttons
don't control volume level.  The mute button mutes, and the volume
buttons unmute, but volume doesn't change.  Also, lowering volume to
minimum turns on the mute indicator in the volume-level pop-up (even
though it's not muted), but the mute button doesn't.  Software sliders
in gnome-volume-control and system-config-soundcard do work.

>  
>  2) Intel wireless - John Linville, our fearless (and handsome) wireless
>  maintainer, just added a fix for iwlwifi that should help some Intel
>  wireless users. 
Almost everything seems OK, except for attaching to a hidden SSID
(wasn't working before, but I can't test till tomorrow) and vpnc
dropping it's connection.  I need to test if that's an issue with wired
connection before I conclude it's iwl3945-related.

>   
>  
>  Both of the above patches are in kernel-2.6.23.1-37.fc8, which should
>  land in rawhide tomorrow. You can get it from koji in the meantime:
>  http://koji.fedoraproject.org/koji/buildinfo?buildID=22511
>
>  ...
>  
>  === NetworkManager bugs ===
>  
>  * Bug 311511: NEW - dcbw
>  - x86_64 iwl3945 NM crash after Keyring loop Deny
>  
>  Comments indicate that the crash is fixed, but it still loops on bad
>  passwords.  
I think that was my comment.  By "loops" I meant that one is prompted
repeatedly for the keyring password.  There's no failure mode other than
getting the password right or clicking "deny".  Not sure if that's by
design or not.

>  
>  
>  * Bug 323371: NEEDINFO - dcbw
>  - NM 0.7 doesn't support WPA/WPA2 Enterprise (certificate-based
>  authentication)
>  
>  Reporter seems to indicate that this is mostly fixed, but some testing
>  on a fresh install is needed to confirm the fix.
>  
>  
>  * Bug 353091: MODIFIED - dcbw
>  - NetworkManager-0.7.0-0.3.svn3020.fc8 crashes when trying to switch to
>  wireless Network
>  
>  nm-applet crasher. A fix will be in the next build, and then this can be
>  retested and (hopefully) closed. 
>  
>  
>  ...
>  
>  And there you have it. Most of those bugs already have fixes pending,
>  and none of the rest are hair-on-fire critical bugs, so I think we're in
>  good shape. If you can help test the stuff I mentioned above, or you
>  have found some huge problem that we've missed, please let us know.
>  
>  Thanks for helping make Fedora more awesome!
>  
>  -w 
-- 
                Matthew Saltzman

Clemson University Math Sciences
mjs AT clemson DOT edu
http://www.math.clemson.edu/~mjs