On Mon, 1 Feb 2021 18:08:38 +0100 Holger Just <[email protected]> wrote:
Hi, first of all, thanks for the detailed report :) > Package: runit-sysv > Version: 2.1.2-25 > > Before the split of runit-sysv from the runit package in stretch, > runit used to signal PID 1 after it has inserted or removed its > configuration in /etc/inittab in runit.postinst resp. runit.postrm. > > This behavior was apparently removed in 001cc301 [1] with the > intention to re-add matching behavior to the new runit-sysv package > later. > > However, only the logic to add the entry to /etc/inittab was > introduced to runit-sysv in 74a58bf6 [2] while sending a SIGHUP to > PID 1 was not added here. > > This results in runit itself not starting after a fresh installation > of the runit-sysv package. Users have to manually run `telinit -q` or > `kill -s HUP 1`. > > This is a regression. The previous behavior to automatically send a > SIGHUP to PID 1 in postinst and postrm should be re-introduced. > > [1] > https://salsa.debian.org/runit-team/runit/-/commit/001cc301ddd184f2645f35e7551f1fbd14c6d7b8 > [2] > https://salsa.debian.org/runit-team/runit/-/commit/74a58bf60e73ebc63bf62391eb092e1edb09013d This package will be scheduled for removal in unstable, because the sysv and systemd integration have been merged into a single package. (see #953875) The bad news is that this bug is not RC, so I think the release team will not allow for a new version of runit in Stable to fix this.. The good news is that I believe it's already fixed in unstable, in the new package runit-run, see https://salsa.debian.org/debian/runit/-/commit/754e00a3f0eb016d809aefc02ceedbc9854b5e68 you may test it by grubbing it from unstable, or just wait until Bullseye is released. If you think this bug still affects runit-run then please follow up here so i can clone this bug and fix it in unstable. Best Regards, Lorenzo
