Re: nfs client 'nobody no matter what' clarification was Re: k5userok .c / doc mismatch

[email protected]
Newsgroups gmane.comp.encryption.kerberos.heimdal.general
Message-ID <[email protected]>
On Fri, Oct 14, 2011 at 10:16:45PM -0500, Harry Coin wrote:
> Now, these daemons are inter-related nfs clients  that update files on 
> all manner of file services and servers.   The whole local world 

Nothing prevents you from allocating account names of any form, like
service-XYZ in the passwd database of the NFS-server(s). Then give the
corresponding NFS-client's service daemon's uid a keytab for principal
named [email protected] and let it fetch the corresponding tgt,
and you are done. The local uid of the daemon is of no interest.

It is sufficient if a separate process with the daemon's uid fetches a TGT
for the [email protected] (and refreshes it). The daemon may remain
fully unaware why it is able to access the NFS areas which it needs.

I suspect you take too many of the exsting practices' limitations
as a reflection of reality. No.

- you do not need to swich the whole OS to a different library if
  you need the corresponding functionality for a certain application

- you do not have to synchronize:
  - numerical gid/uids between different computers
  - principal names with local account names which are using the identities
  - passwd databases between NFSv4 servers and clients
  - the Kerberos identity being used to verify the clients of a service
    and the identity used by the service daemon as another service
    client (say, for NFS' sake)

As one of the consequences, I do not see why you'd need any complicated
mapping between principals and host-local account names.
As another one, you do not have to rely on LDAP, this may relieve
some pains.

Hope this helps somehow.

Regards,
Rune
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.