Hi,
Sorry for the delayed answer
On Sat, Jan 19, 2019 at 4:36 PM Dmitry Bogatov <kact...@debian.org> wrote:
> [2019-01-16 06:32] Niels Thykier <ni...@thykier.net>
> > Felipe Sateler:
> > > Bringing in the debhelper maintainers into the loop again.
> > >
> > > [...]
> > >> Should an init script be executed, invoke-rc.d always returns the
> status
> > > code returned by the init script.
> > >
> > > So it looks like this would be quite the change in behavior.
> > >
> > > I wonder what the best approach it. OTOneH, I sympathize with Joey's
> > > argument that these checks should be centralized. OTOtherH, I'm not
> sure we
> > > can consider invoke-rc.d a debhelper implementation detail and change
> > > behaviour like this.
> > >
> > > Maybe a new flag to invoke-rc.d could be used for this? I'm not sure.
> >
> > I am fine with using a new flag in a new compat level provided it is
> > available in stable at the time. I.e. it would help a lot that the flag
> > is available in buster, so we can add it in the first compat level
> > released in bullseye.
>
> Wonderful. What about this patch:
>
> From 6c856e13dfd496b25652a202f210177dd0f86c19 Mon Sep 17 00:00:00 2001
> From: Dmitry Bogatov <kact...@debian.org>
> Date: Thu, 10 Jan 2019 16:31:32 +0000
> Subject: [PATCH] invoke-rc.d: exit value 6 from init script is fine
>
> Add new option `--lsb' to `invoke-rc.d' script. When this option is
> specified, exit value 6 from init script (service not configured) is
> considered successful invocation (Closes: #629902)
> ---
> man8/invoke-rc.d.rst | 3 +++
> script/invoke-rc.d | 11 ++++++++++-
> 2 files changed, 13 insertions(+), 1 deletion(-)
>
> diff --git a/man8/invoke-rc.d.rst b/man8/invoke-rc.d.rst
> index e5aaeee..53dd435 100644
> --- a/man8/invoke-rc.d.rst
> +++ b/man8/invoke-rc.d.rst
> @@ -88,6 +88,9 @@ OPTIONS
> Return status code 101 instead of status code 0 if
> the init script action is denied by the policy layer.
>
> +*--lsb*
> + Consider exit status 6 (not configured) of init script as success.
> +
> *--query*
> Returns one of the status codes 100-106. Does not
> run the init script, and implies *--disclose-deny*
>
This should also be documented in the STATUS CODES section.
> diff --git a/script/invoke-rc.d b/script/invoke-rc.d
> index 27c045e..6b11fdf 100755
> --- a/script/invoke-rc.d
> +++ b/script/invoke-rc.d
> @@ -34,6 +34,7 @@ ACTION=
> FALLBACK=
> NOFALLBACK=
> FORCE=
> +LSB=
> RETRY=
> RETURNFAILURE=
> RC=
> @@ -77,6 +78,8 @@ Options:
> Return status code 101 instead of status code 0 if
> initscript action is denied by local policy rules or
> runlevel constrains.
> + --lsb
> + Consider exit status 6 (not configured) of init script as success.
> --query
> Returns one of status codes 100-106, does not run
> the initscript. Implies --disclose-deny and --no-fallback.
> @@ -238,6 +241,9 @@ while test $# -gt 0 && test ${state} != III ; do
> --try-anyway)
> RETRY=yes
> ;;
> + --lsb)
> + LSB=yes
> + ;;
> --disclose-deny)
> RETURNFAILURE=yes
> ;;
> @@ -550,9 +556,12 @@ if test x${FORCE} != x || test ${RC} -eq 104 ; then
> elif [ -n "$is_openrc" ]; then
> rc-service "${INITSCRIPTID}" "${saction}" && exit 0
> else
> - "${INITDPREFIX}${INITSCRIPTID}" "${saction}" "$@" && exit 0
> + "${INITDPREFIX}${INITSCRIPTID}" "${saction}" "$@"
> fi
> RC=$?
> + [ "${RC}" = 0 ] && exit 0
> + # service not configured. See #629902
> + [ x"${LSB}" != x ] && [ "${RC}" = 6 ] && exit 0
>
Lets go the other way around: [ x${LSB} = xyes ] , I find that easier to
read. Also, we should probably also allow exit code 5
http://refspecs.linuxbase.org/LSB_3.0.0/LSB-PDA/LSB-PDA/iniscrptact.html
This only works if invoking the init scripts directly. systemctl will hide
the exit status of the init script (the default behavior of systemd is to
accept exit code 5 and 6 for LSB scripts though). No idea what rc-service
does. This should probably be documented in the manpage.
--
Saludos,
Felipe Sateler
