Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Kprop failing when changing permitted tgs and tgt types

142 views
Skip to first unread message

Joe Travaglini

unread,
Feb 3, 2013, 8:35:02 AM2/3/13
to kerb...@mit.edu
Hello,
I had a successfully replicating KDC environment which was working.
I then changed the default and permitted tgs/tgt encryption types,
removing aes256 and aes128 from all fields.

I now get the following error when running kprop after a kdb5_util dump:
kprop: Message size is incompatible with encryption type while encoding
database block starting at 0

I tried destroying and recreating both KDCs to start fresh but I still am
getting this error. After extensive research I cannot understand why. Can
anyone help?

thank you
-Joe

Greg Hudson

unread,
Feb 3, 2013, 11:22:10 AM2/3/13
to Joe Travaglini, kerb...@mit.edu
On 02/03/2013 08:35 AM, Joe Travaglini wrote:
> I tried destroying and recreating both KDCs to start fresh but I still am
> getting this error. After extensive research I cannot understand why. Can
> anyone help?

You found a bug; kprop doesn't work with an RC4 session key. This was
previously unknown to me (or to our bug database, as far as I can tell);
I've written it up here:

http://krbdev.mit.edu/rt/Ticket/Display.html?id=7561&user=guest&pass=guest

For the moment, you'll need to use any other enctype as your preferred
session key enctype when running kprop.

Joe Travaglini

unread,
Feb 3, 2013, 12:36:40 PM2/3/13
to Greg Hudson, kerb...@mit.edu
OK Greg. Thank you for the reply. Is there any way to configure this
specifically for kdb5 dump and kprop and not globally for all principals?

Much obliged,
-Joe

Greg Hudson

unread,
Feb 3, 2013, 12:48:13 PM2/3/13
to Joe Travaglini, kerb...@mit.edu
On 02/03/2013 12:36 PM, Joe Travaglini wrote:
> OK Greg. Thank you for the reply. Is there any way to configure this
> specifically for kdb5 dump and kprop and not globally for all principals?

You could use a separate krb5.conf (and the KRB5_CONFIG environment
variable), but short of that, no.

Why are you removing AES from the enctype variables? Perhaps there is a
better way around whatever problem you're fixing that way.

Joe Travaglini

unread,
Feb 3, 2013, 12:50:59 PM2/3/13
to Greg Hudson, kerb...@mit.edu
It's an application limitation but I'm now considering fixing the
application config given this info. Thanks again, very helpful.
0 new messages