[systemd-devel] F16_64: attempt at OpenVPN server service file
Michael D. Berger
m.d.berger at ieee.org
Mon Nov 28 06:49:03 PST 2011
> -----Original Message-----
> From: Zbigniew Jedrzejewski-Szmek [mailto:zbyszek at in.waw.pl]
> Sent: Monday, November 28, 2011 04:38
> To: Michael D. Berger
> Cc: systemd-devel at lists.freedesktop.org; 'Reindl Harald'
> Subject: Re: [systemd-devel] F16_64: attempt at OpenVPN
> server service file
>
> On 11/27/2011 11:30 PM, Michael D. Berger wrote:
> >
> >> -----Original Message-----
> >> From: Zbigniew Jedrzejewski-Szmek [mailto:zbyszek at in.waw.pl]
> >> Sent: Sunday, November 27, 2011 14:12
> >> To: systemd-devel at lists.freedesktop.org
> >> Cc: Michael D. Berger; 'Reindl Harald'
> >> Subject: Re: [systemd-devel] F16_64: attempt at OpenVPN server
> >> service file
> >>
> >> On 11/26/2011 11:39 PM, Michael D. Berger wrote:
> >>>> than i type "systemctl stop whatever.service"
> >>>>
> >>>> Restart is triggered if they process goes away and in
> the case of
> >>>> "Always" this happens even if the process gives back a
> >> successfull 0
> >>>> like after "killall processname"
> >>>>
> >>> So I gather that Restart is triggered only if the process
> goes away
> >>> ***for reasons other than a stop having been issued***.
> >>> I suggest that the man pages be modified to say that.
> >> Maybe like this?
> >>
> >> -- >8 --
> >> Subject: [PATCH] man: clarify that Restart= is only for the
> >> unexpected cases
> >>
> >> ---
> >> man/systemd.service.xml | 5 +++--
> >> 1 files changed, 3 insertions(+), 2 deletions(-)
> >>
> >> diff --git a/man/systemd.service.xml
> >> b/man/systemd.service.xml index 7b6f12d..7c2ed7e 100644
> >> --- a/man/systemd.service.xml
> >> +++ b/man/systemd.service.xml
> >> @@ -462,8 +462,9 @@
> >> <varlistentry>
> >>
> >> <term><varname>Restart=</varname></term>
> >>
> <listitem><para>Configures whether
> >> the
> >> - main service process shall be
> >> - restarted when it exits.
> Takes one of
> >> + service shall be restarted
> >> if the main
> >> + process exits for reasons
> other than
> >> + the service being stopped.
> >> Takes one of
> >> <option>no</option>,
> >> <option>on-success</option>,
> >> <option>on-failure</option>,
> >> --
> >> 1.7.8.rc3.348.g8fea1
> >>
> >>
> >> __________ NOD32 6664 (20111127) Information __________
> >>
> >> This message was checked by NOD32 antivirus system.
> >> http://www.eset.com
> >>
> >
> > The change shown is not bad, but I might say something like
> "... other
> > than the Stop command being issued..."
> There are other reasons why a unit can be stopped: if the
> unit Requires a second unit that was stopped, if isolate is
> requested, and so on. For me, "a Stop command being issued"
> seems to take into account only the case of this specific
> unit being stopped.
>
> Zbyszek
>
[...]
Excellent point. I might then condider something like:
The circumsiances in which the restart process would be initiated are:
* program failure of the unit
* killall issued on the unit
* killall issued on another unit that this unit Requires (? I am
guessing)
* etc., make the list complete
The circumsiances in which the restart process would NOT be initiated are:
* The Stop commnd is issued to the unit
* The Stop command is issued to another unit that this unit Requires
* etc., make the list complete
I find bullet lists easy to read quickly, but other formats are acceptable
***provided the information given is clear and complete***.
In my long experience as a reader of numerous man pages, often too much
is left to trial and error.
Mike.
--
Michael D. Berger
m.d.berger at ieee.org
http://www.rosemike.net/
More information about the systemd-devel
mailing list