On Thu, Dec 20, 2018 at 1:15 AM Chris Zubrzycki <ch...@mid-earth.net> wrote: > > I would like to amend my reply. The files are no longer in the deb, however > they are still listed by dpkg -L zfsutils-linux. I do not understand how the > solution to an ordering issue is to remove needed init files that have been > fought to be included for literally years. >
The init.d scripts are not present in zfsutils-linux/0.7.12-1~bpo9+1, see: https://packages.debian.org/stretch-backports/amd64/zfsutils-linux/filelist Literal reasons about whether init.d scripts should be included in stretch-backports version do not make sense to such a serious storage package, which may cause interruptions (or data loss in extreme situation) when unexpected conditions occur. I'm temped to not ship init.d scripts for Buster if there is any important issue open when freeze approaches (e.g. bug #915831). It's not a fight for any literal reason about init.d scripts but I don't want to ship something known to be breaking existing stuff, also I wish the package behaves (almost) identically on all supported init (systemd, sysv/insserv, sysv/openrc, if desired). Because I don't have an enterprise environment to battle test these stuff, I can't promise to test and fix issues in non-stable branches. Stable branches I consider are: stable, stable-bpo, oldstable-bpo{,-sloppy}. Let me clarify about what is needed: please anyone interested, step in and take care of init.d scripts by upstreaming any changes needed, also make sure to test them extensively and continuously. I'm happy to apply init.d changes ack'ed by upstream, but will discard any non-trivial feature-wise debian change (i.e. the OpenRC patch). Regards, Aron
