[Freeipa-devel] [PATCH] 0033 Add disconnect command to change topology

[Simo Sorce]

On Tue, 21 Dec 2010 21:57:29 +0100
Jakub Hrozek <jhrozek at redhat.com> wrote:

> On 12/21/2010 07:48 PM, Simo Sorce wrote:
> > On Tue, 21 Dec 2010 18:47:36 +0100
> > Jakub Hrozek<jhrozek at redhat.com>  wrote:
> >
> >> -----BEGIN PGP SIGNED MESSAGE-----
> >> Hash: SHA1
> >>
> >> On 12/21/2010 03:10 PM, Jakub Hrozek wrote:
> >>> On 12/20/2010 09:04 PM, Simo Sorce wrote:
> >>>> On Mon, 20 Dec 2010 18:22:48 +0100
> >>>> Jakub Hrozek<jhrozek at redhat.com>  wrote:
> >>>
> >>>>> -----BEGIN PGP SIGNED MESSAGE-----
> >>>>> Hash: SHA1
> >>>>>
> >>>>> On 12/16/2010 02:02 AM, Simo Sorce wrote:
> >>>>>>
> >>>>>> This command will delete a replication agreement unless it is
> >>>>>> the last one on either server. It is used to change replication
> >>>>>> topology without actually removing any single master for the
> >>>>>> domain (the del command must be used if that the intent).
> >>>>>>
> >>>>>> Simo.
> >>>>>>
> >>>>>
> >>>>> Please document the new action in the manpage. As the actions
> >>>>> are not printed when one specifies --help, there's no way to
> >>>>> discover it short of reading the code.
> >>>
> >>>> I have a separate ticket to add all the changes to the man page.
> >>>> It requires some deep review and I preferred to split it from the
> >>>> rest of the changes.
> >>>
> >>>> Simo.
> >>>
> >>>
> >>> OK, as long as it is tracked I'm fine :-)
> >>>
> >>> Ack
> >>
> >> Actually, sorry, one more thing I noticed. In del_link(), you
> >> assign "type2 = repl2.get_agreement_type(replica1)" but never use
> >> type2 again. Should the next if say "if repl2 and type2 ==
> >> replication.IPA_REPLICA:" perhaps?
> >
> > Initially that was the aim, but then I realized that you wouldn't be
> > able to delete replication agreements with Windows domains
> > (winsync), if you enforced that so I just removed the check about
> > type2 and type2 was left unused.
> >
> > I can remove that line before pushing if needed.
> >
> > Simo.
> >
> 
> Yes, I think that would be nice.
> 
> Ack again and thanks for explaining!

removed the type2 line and also added a repl2 = None on top (it was
causing errors in a later patch).

Pushed to master.

Simo.

-- 
Simo Sorce * Red Hat, Inc * New York