Rawhide Watch

Daily warnings for rawhide victims

Boot failures with kernels installed after grubby-8.31-1.fc21

Posted by adamwill on April 1, 2014

There was a bug in grubby-8.31-1.fc21 which caused it to write invalid grub config entries. This may only affect UEFI installations. If you installed any kernels after installing that version of grubby, they will likely be unbootable. When you try and boot them you’ll see an error like this:

error: file ‘/vmlinuz-3.14.0-0.rc8.git1.1.fc21.x86_64′ not found.
error: can’t find command ‘devicetree’.
error: you need to load the kernel first.

The bug in grubby is fixed with grubby-8.32-1.fc21, but that’s not the whole story. grubby works by using the top entry in the existing grub config file as a ‘template’ for the new entry, so as long as you have a ‘bad’ entry written by 8.31 as your top entry, even the new grubby will still write broken entries. To completely resolve this issue, you’ll need to manually edit your /etc/grub2.cfg and remove all broken entries. You should still have at least one working entry – the entry for whatever kernel is still bootable. Once you remove all broken entries, things should work OK again. You can reinstall kernel packages whose entries were lost, or use grub2-mkconfig to regenerate the grub2 config file entirely.

Posted in Uncategorized | Comments Off

GNOME start failure with new selinux-policy

Posted by adamwill on February 26, 2014

At least for the author, with selinux-policy-3.13.1-27.fc21 , GNOME fails to start correctly if SELinux is in Enforcing mode. If SELinux is in Permissive mode it starts correctly. This denial seems most likely to be the cause of the problem.

Posted in Uncategorized | Comments Off

Long-term users: check /var/run is a symlink to /run before updating to systemd 210

Posted by adamwill on February 26, 2014

If your Rawhide system has been around for a while, check that /var/run is a symlink to /run before you update to systemd 210. If it isn’t, the system will likely fail to boot.

Recent Fedora installations create /var/run as a symlink to /run, but older ones didn’t (I don’t know precisely when this was changed), and it seems systemd and dbus disagree about the location of dbus’ socket: dbus thinks it should be in /var/run , and systemd thinks it should be in /run. If one’s a symlink to the other this is fine; if it isn’t, things are very much not fine.

If /var/run isn’t a symlink to /run, you can move its contents into /run and make it into one. (Even if you mess up, this isn’t terribly important, as all this stuff is supposed to be transient anyway: /run is a tmpfs. Worst you can do is break one boot.)

Bug filed upstream for systemd dbus.

Posted in Uncategorized | Comments Off

Boot failures with recent Rawhide kernels

Posted by adamwill on February 4, 2014

Several people have reported that recent kernels sometimes, often or always completely fail to boot, very early in the boot process (right after the boot menu). We’re still looking into this at present, but we have a possible indication of what the problem might be right now. You may see a trace in numa stuff if you boot with earlyprintk=vga.

We’d advise Rawhide users to make sure you still have a 3.13 kernel installed, just in case you run into trouble booting 3.14 kernels.

Edit: kernel-3.14.0-0.rc1.git0.2.fc21 may well fix this. Please test it out, if you’re affected, and let us know if it doesn’t.

Posted in Uncategorized | 2 Comments »

GNOME fail (again) thanks to libevdev 0.6-1

Posted by adamwill on December 27, 2013

A couple of days ago, a new libevdev build, 0.6-1, was sent to Rawhide. It changed the ABI of libevdev.so.1 by dropping some ‘deprecated’ functions, without bumping the library version (soname). Unfortunately, clutter was still using these ‘deprecated’ functions: upstream clutter git had been updated to use the replacement functions, but no release had yet been cut with this change, and it had not been backported to Fedora.

This meant that if you installed libevdev-0.6-1.fc21, GNOME would fail to start up, with gnome-shell reporting a failure in clutter:

gnome-shell: symbol lookup error: /lib64/libclutter-1.0.so.0: undefined symbol: LIBEVDEV_READ_NORMAL

We’ve now sent out a libevdev-0.6-2.fc21 build which restores the dropped functions, and also a clutter-1.16.2-4.fc21 build which backports the fix to use the replacement functions, thus fixing the bug from both ends for extra safety. If you’re currently suffering from this bug, updating libevdev, clutter or both should resolve it.

Posted in Topical but Not Interesting Later | 1 Comment »

SELinux execmem and execstack denials for everything, and GNOME not running

Posted by adamwill on December 23, 2013

So hey, this blog still exists!

For this author (AdamW), there are two major issues in Rawhide currently. SELinux appears to be denying ‘execmem’ and ‘execstack’ access to all sorts of stuff – this is affecting dhclient (stopping networking working), sshd, sssd (for those of us who use FreeIPA authentication), cups and several other things. I haven’t looked into what caused this exactly yet, but booting with enforcing=0 is likely to be a requirement for Rawhide users until it gets figured out. Bugs filed so far are #1046112 and #1045682.

Also for this author, GNOME currently fails to run at all. GDM doesn’t start, and if I boot to runlevel 3 and run startx I get an extremely broken GNOME session with Shell not running and windows all over the shop. For now, I’ve switched to Xfce. I suspect this is because about half of GNOME 3.11.3 got built before most of the desktop team left for the holidays, but again don’t have precise details yet. We may be able to get it fixed up with the help of any desktop folks who are still around.

Update on GNOME: turns out it just needs a new gsettings-desktop-schemas build. That is now in progress – grab that when it’s done, and GNOME should work again, with all other components at their latest Rawhide level.

Posted in Topical but Not Interesting Later, Uncategorized | 4 Comments »

No input when booting Rawhide of 2010-01-28

Posted by adamwill on January 28, 2010

When booting Rawhide of 2010-01-28, you may hit a tricky situation: absolutely no input at all. In some (possibly all) cases this affects runlevel 3 as well as runlevel 5. This is due to bug #559679: it’s an SELinux problem which prevents HAL starting up (so can cause problems for anything else that still uses HAL, as well).

To fix this problem, you need to install selinux-policy 3.7.8-5 or later (update all selinux-policy packages you have installed, of course). To install them, boot to runlevel 3: add ’3′ to your kernel parameters in grub. If you get no input even at runlevel 3, try deleting ‘rhgb quiet’ from the grub parameters as well. Install the selinux-policy updates, reboot, and input should be working again.

Obviously, if you did not yet update your Rawhide (or you updated but didn’t reboot), make sure when you do update that if you get an older selinux-policy, then install the newer one from Koji before you reboot! That way you’ll avoid hitting the problem.

Posted in Topical but Not Interesting Later | 2 Comments »

Nightly LiveCDs not booting

Posted by adamwill on September 22, 2009

Some people are now aware of (and using) the nightly Rawhide live builds. To save time and bandwidth, it should be noted that all nightly builds since 2009-09-18 fail to boot at all, on any hardware, and are hence essentially unusable. We are currently looking into the cause of this problem and hope to have the nightly builds working again soon.

Update 2009-09-24: The initial problem was resolved, but 2009-09-22 and 2009-09-23 nightly builds were affected by a second problem with the same result. We hope that 2009-09-24 builds will finally fix all issues and boot correctly, once they are available (which depends on when the main Rawhide compose is completed).

In the mean time, the latest available working live build is Snap3, which is available at the pre-release download page. This was built from the 2009-09-17 Rawhide tree.

Posted in Topical but Not Interesting Later | 1 Comment »

restorecond in policycoreutils-2.0.71-6 breaks the world, how to recover

Posted by rawhidewatch on August 20, 2009

If you were running the restorecond service and upgraded to policycoreutils-2.0.71-6.fc12 then things went horribly wrong.  restorecond goes nuts and screws up labels across your filesystem.   Logins via ssh or local console breaks among other problems.

Here is how to recover:

  1. Boot into single user mode.
  2. Turn off restorecond service with chkconfig restorecond off
  3. fixfiles restore will fix labels across your entire filesystem.
  4. Reboot into the normal system.
  5. Upgrade to policycoreutils-2.0.71-7.fc12 or later.
  6. chkconfig restorecond on

In related news, you might be having problems with the restorecon command.  In the past commands like restorecon -Rv . would relabel the current directory, but now it fails silently.  dwalsh says this is a bug he intends to fix.  Until it is fixed, you must specify an absolute pathname to the restorecon command.

Posted in Uncategorized | Comments Off

Rawhide X crashes on startup

Posted by rawhidewatch on August 6, 2009

xorg-x11-server-1.6.99-25* is broken.  Upgrade to -27 to fix this problem.  This is should be pushed to rawhide before Saturday.

Posted in Topical but Not Interesting Later | Comments Off


Get every new post delivered to your Inbox.