Pseudobacklinks patches and related patches

[Stefan (metze) Metzmacher]

Am 12.06.2011 13:20, schrieb Matthieu Patou:
> On 12/06/2011 14:27, Andrew Bartlett wrote:
>> On Sun, 2011-06-12 at 01:04 +0400, Matthieu Patou wrote:
>>> Hello All,
>>>
>>> I'm planning to push patches from this branch in the main tree:
>>> http://git.samba.org/?p=mat/samba.git;a=shortlog;h=refs/heads/pseudobacklinks
>>>
>>>
>>> The pseudobacklink part has been reviewed by Andrew Bartlett, I think
>>> I've addressed all his remarks.
>> http://git.samba.org/?p=mat/samba.git;a=commitdiff;h=1ff18e3793da50666e566700535352df350e7cbb
>>
>>
>> I'm concerned that you still attempt to modify the @ROOTDSE record.  I
>> still think that any entry in @ROOTDSE that turns out to be dynaimic
>> should be fully dynamic, rather than making changes to @ROOTDSE.
>>
>> I'll make a more full review later, but I wanted to mention that part
>> now.
> I've retried to work on making this dynamic it turns out not to be very
> simple as doing a search with a guid at the very beginning (well when we
> are searching for the DN of dsServiceName for instance) is not working
> not because of indexes as I thought but just because the whole samdb is
> not assembled yet and search are directed to sam.ldb instead of the
> correct partition ldb (ie. CN=Configuration,DC=domain,DC=tld).
> 
> There must be a solution but it's might not be obvious, further more I
> would argue that I made this patch with metze and so the idea of
> modifying directly rootdse didn't sound wrong to him (at least at this
> time). Also I think that's important to release this patch at a
> milestone as running the update script is mandatory and it will easier
> to pass this message at a milestone than between two milestones.

Only storing the guid instead of the dn in the dsServiceName attribute
of @ROOTDSE,
is very tricky, as a lot of ldb module init code relies on
samdb_ntds_settings_dn(),
so it's needed before we load the schema.

I think it's much better to have the pseudobacklink handling with the
rename hack
than no support at all.

If someone has an idea how to solve the init chicken and egg problems,
please tell me:-)

metze

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 262 bytes
Desc: OpenPGP digital signature
URL: <http://lists.samba.org/pipermail/samba-technical/attachments/20110620/581ef602/attachment.pgp>