Stopped on prerm during upgrade, not just remove

[Anders Kaseorg]

Binary package hint: network-manager

Before the upstart transition, network-manager’s prerm would only stop the daemon if the package was being removed. Now, the prerm stops the daemon unconditionally, which means the network is brought down for the entire duration of any upgrade that includes a new network-manager.

This would be fixed by setting DEB_DH_INSTALLINIT_ARGS = -R in debian/rules, except that the transition doesn’t quite work: the old prerm still stops network-manager, and the new postinst tries to restart it, but restarting a stopped Upstart job apparently fails (“restart: Unknown instance: ”).

[Anders Kaseorg (andersk)]

I filed the transition issue with ‘restart’ failing as bug #430883.

[Launchpad Janitor (janitor)]

This bug was fixed in the package network-manager - 0.8~a~git.20090911t130220.4c77fa0-0ubuntu6

---------------
network-manager (0.8~a~git.20090911t130220.4c77fa0-0ubuntu6) karmic; urgency=low

  LP: #430878
  * debian/rules: We never previously started or restarted network manager
    on install or upgrade, and instead flagged a reboot. So we need a
    --noscripts for dh_installinit to stop that happening with Upstart
    (didn't need this before because the init script was never installed
     with dh_installinit)
  * debian/network-manager.prerm: Restore, just stop N-M using "stop"

 -- Scott James Remnant <email address hidden> Wed, 16 Sep 2009 20:50:31 +0100

[Anders Kaseorg (andersk)]

Other packages that now stop unconditionally on prerm are acpid, apport, at, avahi-daemon, cron, hal, ifupdown (network-interface and network), util-linux (hwclock and hwclock-save). I bet most or all of these are bugs.

[Scott James Remnant (Canonical) (canonical-scott)]

On Thu, 2009-09-17 at 22:09 +0000, Anders Kaseorg wrote:

> Other packages that now stop unconditionally on prerm are acpid, apport,
> at, avahi-daemon, cron, hal, ifupdown (network-interface and network),
> util-linux (hwclock and hwclock-save). I bet most or all of these are
> bugs.
>
Did you check to see whether they unconditionally stopped before?
That's *always* been default behaviour.

If not, could you do that please?

Scott
--
Scott James Remnant
<email address hidden>

[Anders Kaseorg (andersk)]

> Did you check to see whether they unconditionally stopped before?
> That's *always* been default behaviour.

Hmm, yeah, acpid, apport, at, avahi-daemon, cron, and hal were unconditionally stopped before. But the other two packages did change behavior:

• The jobs in ifupdown were previously udev rules, and ifupdown did not previously have a prerm script at all.
• util-linux previously did not stop /etc/init.d/hwclock.sh on prerm.

[Launchpad Janitor (janitor)]

This bug was fixed in the package ifupdown - 0.6.8ubuntu21

---------------
ifupdown (0.6.8ubuntu21) karmic; urgency=low

  * --no-scripts is not --noscripts. LP: #430878.

 -- Scott James Remnant <email address hidden> Tue, 22 Sep 2009 15:30:28 -0700

[Launchpad Janitor (janitor)]

This bug was fixed in the package util-linux - 2.16-1ubuntu4

---------------
util-linux (2.16-1ubuntu4) karmic; urgency=low

  * Removed access checks for hardware clock when called with --systz,
    since we may not have the rtc device at the point we run hwclock.
    I believe this is the cause of LP: #436076.
  * Set kernel timezone even when the hardware clock is in UTC.
    LP: #426886.

  * Don't step the system clock, or save the hardware clock on upgrades
    in case the time isn't quite correct. LP: #430878.
  * Remove the hwclock.sh and hwclockfirst.sh scripts on upgrades, since
    these are now Upstart jobs. LP: #434767.

 -- Scott James Remnant <email address hidden> Thu, 24 Sep 2009 12:31:29 -0700