[Debian-ha-maintainers] Pacemaker packaging

[Richard Winters]

Hello,

On Sep 11, 2015 10:39 AM, "Ferenc Wagner" <wferi at niif.hu> wrote:
>
> Hi,
>
> I almost went through the packaging of Pacemaker 1.1.13.  A couple of
> questions arose:
>
> 1. Richard put in Lintian overrides like this:
>
>    # init.d scripts not registered with update-rc.d since other
components are meant to start/stop pacemaker
>    # and friends at system boot
>    pacemaker-remote binary:
script-in-etc-init.d-not-registered-via-update-rc.d
>
>    What are those "other components"?  What happens under systemd?
>

In the past cman, in the present pcs...

> 2. Richard also put in Lintian overrides like this:
>
>    # False positive; is as intended by upstream
>    pkg-config-bad-directive
>
>    Richard, could you please point us to your communications with
>    upstream forming the base of the above comment?  Declarations like
>
>    with_upstart=@SUPPORT_UPSTART@
>    with_systemd=@SUPPORT_SYSTEMD@
>
>    don't seem particularly useful to me.

It was discussed in IRC and maybe on github, I'll check issues for a
reference later on.

> 3. Do we want to provide numbered dev packages, like liblrmd1-dev?  I
>    don't think we want to support several API versions concurrently,

I think so, so dependencies all around are properly resolved as best they
can be.

so
>    I dropped the numbers from the dev package names for now.

OK.

> 4. Lintian is not happy about the init script for pacemaker_remote
>    providing pacemaker_remoted (note the final d).  Should we just
>    remove the d and thus synchronize to pacemaker_remote.service?

Why does lintian have a problem with the letter d?  I don't think that's
necessary, but I couldn't say whether it'd hurt anything.

> 5. I dropped the CTS stuff from here as well, just like from Corosync.
>    This made the pacemaker-dev (probably long misnamed) package
>    disappear with all the Python stuff, too.  Objections?

Just the question of why?  Did it hurt anything at all?

> More will come later, but let's start with the above...

OK...

> --
> Thanks,
> Feri.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/debian-ha-maintainers/attachments/20150911/b9f6b319/attachment.html>