Passdb and smbpasswd (following discussion in TODO list thread)

Kevin Colby kevinc at grainsystems.com
Tue Oct 3 14:27:53 GMT 2000


(1) & (2) are both side effects of dynamic RID generation.
The UID-based RID vs. stored RID issue has come up before.
Wasn't SURS supposed to do something about this?
Is that still happening?

Personally, it seems that stored RIDs are absolutely
necessary in the future, mostly for (1) and (2), actually.

	- Kevin Colby
	  kevinc at grainsystems.com


Simo Sorce wrote:
> 
> On the way of discussion all focused on PAM with few point outs to Ldap.
> 
> As not all the OS have PAM and not all wnat to use LDAP I want to
> address the current way information are stored on smbpasswd file.
> 
> What I found annoyng in smbpasswd file structure is:
> 1. The need to store workstation accounts in system passwd.
>    This is annoying as force the administrator to change custom
>    scripts to deal with this entrys in passwd, and as samba does
>    not need to perform any operation under the uid of wks account
>    their presence in passwd is totally useless (and unelegant).
> 2. The fact that RIDs are not stored anywhere.
>    Storing RIDs on smbpasswd may allow an easyer migration path
>    from existing NT domains and allow admins to set particular
>    RIDs if needed.
> 3. Trusting ACB bits User or Group belongings may be discovered
>    and also Workstation accounts may be recognized.
> 
> I made such a patch for samba 2.0.7 version.
> The patch avoided the need to store workstation accounts in passwd
> and also placed the RID field in smbpasswd file.
> If you think this may fit in samba I may change the patch to samba 2.2
> or HEAD as you wish.
> 
> Simo.
> 
> --
> Simo Sorce - Integrazione Sistemi Unix/Windows - Politecnico di Milano
> E-mail: simo.sorce at polimi.it
> Tel.int: 02 2399 2425 - Fax.int. 02 2399 2451
> -----------------------------------------------------------------
> Be happy, use Linux!




More information about the samba-technical mailing list