Small fix to idmap_set_mapping

[Volker Lendecke]

On Tue, Jul 01, 2003 at 11:42:14AM +0200, Simo Sorce wrote:

> Uhmmm. yes... we need to address the problem, probably the best thing is
> to detect the failure in the init_wellknownsid funtion and try with a
> local_only_mapping (add a special idmap flag if necessary).

I just successfully got rid of one :-)

And I still think that simply not trying to store anything that is
already there in this exact form doesn't really cost us much and deals
with the init_wellknownsids quite nicely. On a BDC the mappings should
already be there even upon the first startup. The only thing that then
happens is that the local cache is filled. We don't have to bother the
PDC's LDAP server upon each startup of a BDC.

> The previous code had not this problem as the remote mapping failure was
> a possible outcome, and this was right one of the situation that made me
> think of that way to make the code, too bad I didn't put comments at the
> time :-/

Local success and remote failure leads to inconsistent databases. How
did you want to cope with that?

Volker
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 232 bytes
Desc: not available
Url : http://lists.samba.org/archive/samba-technical/attachments/20030701/df54eda2/attachment.bin