On Thu, Jan 02, 2020 at 09:39:38AM -0800, Adam Williamson wrote: > On Thu, 2020-01-02 at 11:18 -0600, Richard Shaw wrote: > > On Thu, Jan 2, 2020 at 10:14 AM Zbigniew Jędrzejewski-Szmek < > > zbys...@in.waw.pl> wrote: > > > > > On Thu, Jan 02, 2020 at 10:45:07AM -0500, Neal Gompa wrote: > > > > On Thu, Jan 2, 2020 at 10:11 AM Ben Cotton <bcot...@redhat.com> wrote: > > > > Even with this proposal, how do we deal with system upgrades? In the > > > > system upgrade case, we don't really want these restarts to run, > > > > because the system will be rebooted anyway... These kinds of questions > > > > are why I didn't propose porting over what we did in OpenMandriva to > > > > Fedora. It's rather wasteful in a rather common case... > > > > > > I'm not sure if skipping restarts on upgrade is worth the trouble. It > > > should be quite simple to implement (e.g. by having the restart scriptlet > > > check if the systemd target has been diverted to system-update.target), > > > but restarting of services would most likely take only a small fraction > > > of time required for the upgrade anyway. > > > > > > > Maybe it should only restart the service if it's already running. In the > > case of an upgrade it's a special target, right? > > > > Actually that's probably a good idea anyway. If I have stopped a service > > (even if it's enabled) then it should not be restarted on upgrade. > > This is exactly what 'try-restart' does: it restarts the service only > if it's running. Per the docs: > > try-restart PATTERN... > Stop and then start one or more units specified on the command > line if the units are running. This does > nothing if units are not running. > > The current %systemd_postun_with_restart scriptlet uses try-restart: > > [adamw@adam koji ((koji-1.19.1))]$ rpm --eval "%systemd_postun_with_restart > foo" > > > if [ $1 -ge 1 ] && [ -x /usr/bin/systemctl ] ; then > # Package upgrade, not uninstall > /usr/bin/systemctl try-restart foo || : > fi > > The Change page doesn't *specifically* state what scriptlets > it's intended to replace, but I'd assume it would be that one. Thus I'd > expect that this mechanism would also use try-restart.
Yes, exactly. Thank you for the very nice summary. More generally, the goal is to do the exact same restarts that are done now, but at a different point in the transaction and from a different call site. Zbyszek _______________________________________________ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
