On Wed, 25 May 2011 19:14:16 +0200 Michael Biebl <bi...@debian.org> wrote:
> Thanks, this is a dup. You can work around this problem by purging the
> nfs-common/portmap packages.
Ran 'aptitude purge nfs-common portmap' as advised, rebooted.
Still no improvement:
# as 'root'
% dglob nfs-common || dglob portmap ; echo $?
1
% systemadm ; echo $?
(systemadm:24653): GLib-GObject-CRITICAL **: g_object_unref: assertion
`G_IS_OBJECT (object)' failed
(systemadm:24653): GLib-GObject-CRITICAL **: g_object_ref_sink:
assertion `G_IS_OBJECT (object)' failed
0
Attached is the output of:
dmesg | nl | grep -C 5 "systemd.*cycle" | gzip -9 >
/tmp/dmesg_systemd_cycle.log.gz
Then I read more about bug #622881. Parsing the above 'dmesg' extract
for package names gives:
% dmesg | nl | grep "systemd.*Breaking.*cycle" | sed 's,.*job
\(.*\)\..*,\1,'
dbus
avahi-daemon
setserial
ipmasq
A test run of 'aptitude purge dbus avahi-daemon setserial ipmasq' shows it
would need to delete 327 packages. Ouch.
Following your lead, I'm guessing that the above four packages, and the
two before, i.e. 'dbus', 'avahi-daemon', 'setserial', 'ipmasq',
'nfs-common', & 'portmap'; these six packages presumably have buggy
Debian boot dependency code.
HTH...
dmesg_full_systemd_cycle.log.gz
Description: Binary data
