On Mon, Dec 22, 2014 at 05:20:01PM +0000, Tom Hughes wrote:
On 22/12/14 17:07, Jan Kratochvil wrote:
>On Mon, 22 Dec 2014 05:59:50 +0100, Felix Miata wrote:
>>I started a yum upgrade process. When it reached 342/784 (@avahi) over half
>>an hour ago, the screen writing from the process simply halted.
>
>During F20->F21 upgrade I had to run along something like
> while sleep 1;do killall systemctl;done
>(the real command was longer) as many %post scripts tried to restart their
>daemon but during the upgrade each systemctl command was timing out with very
>long delay.
I had that as well - the cause is that systemd asserts and goes into
panic mode which means it no longer responds to systemctl calls. It
didn't happen to all my machines but it did to a couple of them.
The fix is to kill the dbus-daemon process - after that the
systemctl calls will still fail but will do so quickly rather than
slowly.
You will also won't be able to do a clean reboot so will have to
resort to something like "sync; reboot -ff" to reboot after it
completes.
FWIW I saw pretty much identical symptoms on 3 different machines.
Rich.
--
Richard Jones, Virtualization Group, Red Hat
http://people.redhat.com/~rjones
Read my programming and virtualization blog:
http://rwmj.wordpress.com
libguestfs lets you edit virtual machines. Supports shell scripting,
bindings from many languages.
http://libguestfs.org