On Fri, Jan 10, 2014 at 02:32:48AM -0800, Chris Gray wrote:Please note that those options have a special purpose and should not be
> I'll install the ldb-tools tomorrow (I went home) and try that out.
> The SID and the RID are correct, verified visually and used a windows
> utility to search the domain based on the SID to verify it was correct and
> returning the correct account from AD (psgetsid.exe). I'm also working on
> converting the SID and RID to reverse hex so I can use ldapsearch on the
> linux machine to triple verify, but I haven't completed that yet.
> The computers with SSSD version 1.9 correctly show the RID as the last
> digits of the unix UID. My default domain group is Domain Users as well,
> which always has a RID of 0513, and that correctly shows as the last 4
> digits unix GID on the computers with 1.9.
> Reading those bug reports more may had lead to a solution.
> ldap_idmap_default_domain_sid (string)
> Specify the domain SID of the default domain. This will
> guarantee that this domain will always be assigned to slice
> zero in the ID map, bypassing the murmurhash algorithm
> described above.
> Default: not set
> ldap_idmap_default_domain (string)
> Specify the name of the default domain.
> Default: not set
needed in your setup. Nevertheless they might lead to a working
solution by hiding the original issue. With those two option a domain is
pre-created in the idmapping code. If the SID matches the domain SID of
your user the user will get a POSIX ID and you won't see the errors you
posted earlier. But in general the AD provider is able to auto-detect all
domain in your forest and create the needed idmapping entries
automatically. I assume that there is an issue to detect of domain and
its domain SID or to create the needed idmapping entries. This is why I
asekd for the full logs.
> I'll try those out tomorrow as well. I'm not sure they'll work since I
> got them from version 1.9 docs. I don't have them set on the computers
> I have that use SSSD 1.9, and they don't exhibit the problem.
> Thanks again,
> sssd-users mailing list
sssd-users mailing list