severity 762528 minor tag 762528 upstream fixed-upstream thanks
Hi,
On 09/23/2014 08:17 AM, Paul Wise wrote:
> Package: needrestart
> Severity: wishlist
>
> I noticed when running checkrestart and needrestart at the same time
> that needrestart doesn't suggest restarting systemd when it has been
> upgraded. The way to restart systemd is to run this command as root:
>
> systemctl daemon-reexec
while upgrading systemd it should restart itself - shouldn't it? I've
tried to run `apt-get install --reinstall systemd` (systemd 208-8 on
jessie). The .postinst script already does a daemon-reexec -
checkrestart nor needrestart suggest to restart systemd.
If a library is upgraded which is used by systemd than needrestart does
fail:
[main] #1 exe => /lib/systemd/systemd
[main] #1 unexpected cgroup '/'
=> did not find a unit name in /proc/1/cgroup
Failed to issue method call: No unit for PID 1 is loaded.
=> systemctl show 1 did not work, either
[main] #1 running /etc/needrestart/hook.d/10-dpkg
[main] #1 package: systemd
=> dpkg -S /lib/systemd/systemd did found the systemd package, but it
does not contain a Sys-V script:
$ dpkg -L systemd |grep init.d/
Needrestart did not find any way howto restart pid #1 :-( . I've changed
upstream to handle systemd restarts as suggested. Seems to work as expected:
# ./needrestart -v -r l
[Core] Using UI 'NeedRestart::UI::stdio'...
[main] detected systemd
[main] #1 uses non-existing /lib/x86_64-linux-gnu/libattr.so.1.1.0.dpkg-new
[main] #1 is not a child
[..]
[main] #1 exe => /lib/systemd/systemd
[main] #1 is the systemd manager
[..]
Services to be restarted:
systemd manager
[..]
The pseudo service name 'systemd manager' does trigger to use `systemctl
daemon-reexec` instead of `systemctl restart @UnitNames`.
Thanks for reporting,
Thomas
--
:: WWW: http://fiasko-nw.net/~thomas/ ::
::: Jabber: xmpp:tho...@jabber.fiasko-nw.net :::
:: flickr: http://www.flickr.com/photos/laugufe/ ::
signature.asc
Description: OpenPGP digital signature
