Thanks to those who responded to my previous problem.  It seems the
   AFS rsh and rcp need to be SUID root.  Ahem.

   Now, by all means correct me if I am wrong, but isn't the whole
   point of having a kerberized in.rlogind, in.rexecd, rsh, login, and rcp
   that authentication will be passed along?  Maybe not the WHOLE point,
   but surely a large one.

   The following scenario shows hosts manos and cn10, both of which have
   the AFS-provided rlogind, login, ftpd, inetd, rexecd.  The new AFS inetd
   on cn10 has been started in place of the old HP one.

---------------------------------------------------------------------------
~ : manos 1:44pm > klog
Password:
~ : manos 1:44pm > tokens

Tokens held by the Cache Manager:

User's (AFS ID 30379) tokens for [EMAIL PROTECTED] [Expires Dec  1 15:10]
   --End of list--
~ : manos 1:44pm > which rlogin
/usr/ucb/rlogin
~ : manos 1:44pm > rlogin cn10
AFS (R) 3.3 Login
~ : cn10 1:44pm > /usr/ciesin/afs/bin/tokens

Tokens held by the Cache Manager:

   --End of list--
~ : cn10 1:44pm > 
---------------------------------------------------------------------------

   Am I missing something here?  Am I mistaken as to my thoughts of what
   the purpose of these AFS utilities is?  Do others experience this same
   thing?  I've checked and re-checked to be sure the AFS versions of those
   binaries are the ones which are running.

-------
Jeff Blaine
CIESIN Operations

Reply via email to