[PATCH] python indent bugfix in dns_hub.py

[Jeremy Allison]

On Wed, Jan 23, 2019 at 07:28:08PM +0100, Stefan Metzmacher via samba-technical wrote:
> Am 23.01.19 um 19:25 schrieb Jeremy Allison via samba-technical:
> > On Wed, Jan 23, 2019 at 07:00:52PM +0100, Stefan Metzmacher wrote:
> >>
> >> I had the following problem a few times, that
> >> the join itself worked fine, but the 'samba-tool drs replicate' commands
> >> fail after we started the new vampire_dc (more details at the end)
> >>
> >> I found a reproducer for this problem:
> >>
> >> In provision_vampire_dc() I added:
> >>    $ctx->{samba_dnsupdate} = "/bin/true";
> >> after the call to provision_raw_prepare().
> >>
> >> The core problem is that 'samba_dnsupdate' sometimes fails to run
> >> within 20 seconds. I see
> >> "dnsupdate_nameupdate_done: Failed DNS update with exit code 110",
> >> where 110 is ETIMED.
> >>
> >> samba-tool drs replicate needs to resolve the
> >> $NTDSGUID._msdcs.$DNSDOMAIN name in order to work, which may fail.
> >>
> >> samba-tool domain join already adds that name to the database of the
> >> used server and replicates this back before it finishes.
> >> But we're using the RESOLV_WRAPPER_HOSTS file instead of real dns
> >> and samba_dnsupdate with --use-file.
> >>
> >> I think we need to call samba_dnsupdate explicit after the join,
> >> if we make use of --use-file in order to avoid the problem.
> >>
> >> In the long run we need to remove the RESOLV_WRAPPER_HOSTS usage.
> > 
> > Oh fantastic catch ! I must confess I don't understand the interaction between
> > the dns updates and RESOLV_WRAPPER_HOSTS *at all* :-) :-).
> 
> The same applied to me a few hours ago...

And that's why you're smarter than I am metze :-) :-).