Rawhide Watch

Daily warnings for rawhide victims

Rawhide 2009/08/05: xulrunner and python-urlgrabber

Posted by adamwill on August 5, 2009

There are two problems in today’s Rawhide. First, xulrunner has been updated, but Firefox (and other Gecko-based applications) have not been rebuilt. If you update to today’s Rawhide and include the updated xulrunner, Firefox (and other Gecko-based applications applications) will not work. If you haven’t updated yet, you can avoid this problem by excluding xulrunner from your update: ‘yum update -x *xulrunner*’. Remember to take this out again once rebuilt Firefox and other packages have gone into Rawhide. If you already got hit by this problem, you need update to the rebuilt Firefox package, but as of the time of this writing, the build has not got through. You can attempt to manually revert xulrunner to its old version, as an alternative.

Secondly, you might find that yum seems unable to download packages properly, getting partway through, failing with a HTTP Error 200 message, trying the next mirror, and failing again. To fix this, update the python-urlgrabber package: ‘yum upgrade python-urlgrabber’.

Posted in Topical but Not Interesting Later | Comments Off

Blank notification area (systray) workaround

Posted by rawhidewatch on August 3, 2009

Many are experiencing a problem where the notification area (a.k.a. systray, or that place icons like nm-applet appear) being blank in rawhide.  The current workaround is to restart notification-area-applet a few times.  There is a bug report for this issue. Hopefully this will be fixed soon.  Thanks to mclasen for this tip.

Posted in Uncategorized | Comments Off

Unable to update to rawhide – rpmlib(PayloadIsXz)

Posted by rawhidewatch on July 31, 2009

jlaska wrote:
Trying to update your Fedora 11 (or older) system to the latest rawhide content?  You might notice a failure similar to below when attempting to upgrade using yum:

rpmlib(PayloadIsXz) <= 5.2-1 is needed by kernel-PAE-2.6.31-0.107.rc4.git3.fc12.i686 
rpmlib(PayloadIsXz) <= 5.2-1 is needed by kernel-firmware-2.6.31-0.107.rc4.git3.fc12.noarch

Turns out that rawhide packages now use XZ (the new LZMA format) as the default payload compression format.  The details of this new feature can be found on the XZRpmPayloads feature page.  Testers that install directly from Fedora 12 or rawhide should not experience problems.  However, folks upgrading from Fedora 11 or from an older rawhide may need to manually workaround this issue using the following commands:

yum --enablerepo=updates-testing update rpm

Once completed, you may continue with the rawhide upgrade:

yum --disablerepo=* --enablerepo=rawhide update

More information on upgrading to rawhide can be found on the wiki.

Happy testing!

Posted in Uncategorized | Comments Off

Rawhide x86_64 DNS broken? nss-mdns multilib issue

Posted by rawhidewatch on July 9, 2009

Tip of the hat to Will Woods for this tip.

If you're using an x86_64 system (F11 or Rawhide) and having trouble
with DNS (hostname lookups) after recent updates, check to see if you
have nss-mdns.i586 installed *without* nss-mdns.x86_64:

  rpm -q nss-mdns

If so, you should be able to fix name resolution by doing either:

  mv /etc/nsswitch.conf /etc/nsswitch.conf.saved
  mv /etc/nsswitch.conf.bak /etc/nsswitch.conf
  yum install nss-mdns.x86_64

or simply: yum erase nss-mdns

The cause of the problem is a recent change to WINE that made it require
nss-mdns[1]. Since wine is i586-only, the i586 nss-mdns package gets
installed. This is a long-standing bug with multilib support[2] that the
RPM developers are working on. 

[1] https://bugzilla.redhat.com/show_bug.cgi?id=492700
[2] https://bugzilla.redhat.com/show_bug.cgi?id=442047

Posted in Uncategorized | Comments Off

ssh key-based login to root not working?

Posted by rawhidewatch on June 1, 2009

Is your ssh key-based login to root not working? Solution:

  1. chmod 700 /root/.ssh/
  2. chmod 600 /root/.ssh/authorized_keys
  3. restorecon -R -v /root/

Posted in Uncategorized | Comments Off

yum-presto Delta RPM support needs testing

Posted by rawhidewatch on April 20, 2009

yum-presto in rawhide now is in need of widespread testing.  Presto allows yum to download Delta RPM .drpm files, which contain only the difference between your installed package and the upgraded new version.  This allows yum to update packages with significant download bandwidth and time savings.

Please help us verify that yum-presto is working properly.  Please test it with both standard yum command line and PackageKit.  Simply run yum install yum-presto.  No further configurations are necessary.  Please report any problems to Bugzilla.

Posted in Uncategorized | Comments Off

GNOME > System > Shutdown fails to Shutdown?

Posted by rawhidewatch on April 17, 2009

In the past week I’ve noticed the following strange behavior on multiple Rawhide machines.

  • I am logged into the GNOME desktop.
  • I select System > Shutdown > Shutdown
  • X dies, the shutdown animation happens, but it just stops.
  • Ctrl-Alt-F2 switches to a VT, but typing does nothing because the getty is dead.
  • Ctrl-Alt-Delete starts the reboot process, where we see things actually go through the shutdown process with services stopping.

Shutdown from GDM, or shutdown immediately after logging into GNOME do not exhibit this behavior.  I am told that Bug #495326 is related to this issue.

Posted in Uncategorized | Comments Off

ext4 on Intel X25-M SSD needs nodelalloc

Posted by rawhidewatch on April 13, 2009

Many early SSD owners paid hundreds of dollars for a solid state drive disk, only to realize that performance is terrible with “stuttering” caused by terrible SSD controllers made by JMicron and its inability to handle random write patterns.  Intel’s X25-M and 25-E drives have consistently beat all other consumer grade SSD drives due to its superbly designed SSD controller, but at a huge price premium.  Generally users of Intel’s SSD’s have avoided the “stuttering” problems of lesser quality drives.

Unfortunately, it seems that 15-60 seconds of seemingly deadlock-like non-responsiveness can happen even with an Intel X25-M when using ext4 filesystem.  Ever since I upgraded my Thinkpad T60 laptop to an Intel X25-M 160GB drive, my entire system would feel like it freezed, but then I notice the hard drive light flashing intermittently.  After maybe 30-60 seconds the hard drive light would be solid on for a few seconds, then the system would recover.  I cannot figure out how to reproduce this on-demand, but it seems to happen 2-3 times a day, and only during times where disk activity is very low.

In any case, this problem goes away entirely if I add “nodelalloc” to the mount flags of my ext4 filesystems.  This turns off ext4 delayed allocation, which can typically delay writing to the disk for longer periods than ext3′s 5 seconds, perhaps 30-60 seconds.  This is a benefit to hard drives because you can save power by spinning up less often.  But for Intel”s X25-M drives, you are better off writing as early and often as possible since the drive doesn’t spin (no power benefit), and it reportedly handles parallelism internally very well, meaning the OS shouldn’t worry about it because there is no performance benefit.

In related news, it was suggested to me that elevator=noop is useful in your kernel cmdline in grub.conf if your disks are SSD.  While I am personally trying it, I am uncertain if this is a good thing.  While the elevator normally wouldn’t be necessary with a theoretically perfect solid state disk with constant-time random access to any address,  SSD’s are not perfect with constant time access.  Especially in the case of writing many small changes to non-contiguous parts of the disk, SSD’s suffer from what is known as write amplification because SSD’s need to read, erase and re-write entire blocks if you change only a tiny portion of that block.  I wonder if write-combining that happens in the elevator prevents a few I/O writes to non-adjacent but nearby blocks from triggering multiple write events to the same SSD block.  If it turns out the elevator does help to minimize write amplification before it gets to the SSD’s controller, then it would be both a performance benefit as well as helping the longevity of the disk.  I have no clue how to measure this.  Perhaps we need Intel engineers + people who know filesystems to answer this.

In any case, this ext4 behavior with Intel X25-M seems to be a bug in ext4?  I don’t know.

UPDATE:
Jeff Moyer pointed out that recent versions of the CFQ elevator has smart SSD detection.  If SSD is detected, it disables idling during random access reads because the penalty for reading is not severe.  He thinks you are better off using the default CFQ scheduler with these drives.  Jeff also pointed out that elevator=noop will also do write combining.

UPDATE April 17th,2009:
Intel released a firmware update for their X25-M and X18-M SSD’s that improves performance in the worst case scenario.  I flashed my X25-M and turned off nodelalloc to see if this makes a difference.  It seems ext4 still behaves badly without nodelalloc with occasional episodes of 15-60 seconds of unresponsiveness.

Posted in Uncategorized | Comments Off

Video with weird delays?

Posted by rawhidewatch on April 10, 2009

F-11 rawhide users might have noticed weird behaviors with video playing in the past months.  Playing videos in Youtube, mplayer or totem would often play with audio immediately, but you see no video for a few seconds.  Then suddenly many video frames rapidly play, then video becomes smooth in sync with the audio.  It turns out that this was actually problems with video syncing to the audio clock of pulseaudio.  The latest build pulseaudio-0.9.15-9+ substantially improves this weird problem at the beginning of videos, and seems to make video playback smoother as well.  Note that if you are using mplayer, you need this patch to be applied to remove a workaround hack, otherwise you might not notice any improvement.  Hopefully distributors of mplayer will patch soon.

Posted in Uncategorized | 1 Comment »

Rawhide x86_64 firefox and rpm broke? Workaround Procedure

Posted by rawhidewatch on April 8, 2009

Some rawhide x86_64 users have hit a problem where firefox and rpm were broken by a dependency problem.  Adam Williamson wrote:

https://bugzilla.redhat.com/show_bug.cgi?id=494122

The nss-3.12.2.99.3-5 package introduced a dependency on nss-softokn-freebl. However, this dependency was not architecture specific: it could be satisfied by the i586 or x86-64 package. Depending on what packages they had installed, x86-64 users could wind up with only the i586 nss-softokn-freebl package, which broke nss, and – in consequence – at least Firefox and also rpm.

The nss-3.12.2.99.3-7 update fixes this by making the dependency arch-specific. If you have been affected by the issue – you’ll know if
you’re running x86-64 Rawhide and your Firefox and rpm don’t work, and  you can verify by checking what nss-softokn-freebl packages are installed – you can get out of the jam by downloading the x86-64 nss, nss-tools and nss-softokn-freebl packages from a Rawhide mirror and running:

cd /
rpm2cpio PATH_TO_RPM/nss-3.12.2.99.3-7.fc11.`uname -m`.rpm | cpio -i
rpm2cpio PATH_TO_RPM/nss-softokn-freebl-3.12.2.99.3-7.fc11.`uname -m`.rpm | cpio -i
rpm2cpio PATH_TO_RPM/nss-tools-3.12.2.99.3-7.fc11.`uname -m`.rpm | cpio -i

Posted in Uncategorized | Comments Off

 
Follow

Get every new post delivered to your Inbox.