Davor Ocelic <[EMAIL PROTECTED]> writes:
>>   2. Do we really need separate identities for separate machines?  My
>>      understanding was that the whole service/[EMAIL PROTECTED] convention
>>      was for preventing man-in-the-middle attacks for things like
>>      kerberized telnet.  If we're only using kerberos in order to
>>      support AFS, I don't know if it's really necessary.
>
> We use kerberos for everything?

I'm having trouble parsing this.

We don't use kerberos for: http, telnet, pop3/imap, or (AFAICT) any
TCP-based protocol.  As far as I can tell, we use it only for PAM and
AFS.  I'm okay with it staying this way; anything that needs to check
user passwords should be using PAM.


>>      Also, if we go with a single user per service across all hosts,
>>      we can give it the same userid in /etc/passwd and pts,
>>      eliminating the ID>1000 stuff.
>
> How would that affect Debian postinst scripts that do something like 
> 'adduser --system mysql' ?

When debian postinst scripts create users, we would need to pass the
appropriate "-id" flag to pts when creating the corresponding user.
But since debian creates 10<ID<1000 and we've left pts identities
10..1000 unused, there will be no problem.


> However I'd like to see as much simplification as possible, so if you
> don't think we'll ever need this functionality, ... And you cou could also 
> clarify the note about kerberos from above..

Definately.

  - a


-- 
PGP/GPG: 5C9F F366 C9CF 2145 E770  B1B8 EFB1 462D A146 C380


_______________________________________________
HCoop-SysAdmin mailing list
[email protected]
http://hcoop.net/cgi-bin/mailman/listinfo/hcoop-sysadmin

Reply via email to