principal~.kadm5 & C.

Giuseppe Mazza g.mazza at imperial.ac.uk
Tue Jul 22 12:29:59 EDT 2014


Hi Greg,

I have got a good news and a bad one.

On 17/07/14 17:49, Greg Hudson wrote:
> On 07/17/2014 08:59 AM, Giuseppe Mazza wrote:
>> > What do you think? Do you need more info?
> I think I do need more info.  This helps narrow things down, but there
> are still questions:
> 
> * Is the process actually stuck within that krb5_db_put_principal call,
> or is it somehow in a loop doing put_principal calls over and over again?

<<Bad news>>
Each time I run the command below:
root at tt-u1404:~/boo# kdb5_util load kdb-dump

I get the following situation:
root at tt-u1404:/var/lib/krb5kdc# ls -larth
...
-rw-------  1 root root 8.0K Jul 21 16:45 principal~.kadm5
-rw-------  1 root root    0 Jul 21 16:45 principal~.kadm5.lock
drwx------  7 root root 4.0K Jul 21 16:45 .
-rw-------  1 root root 508K Jul 21 16:45 principal~
-rw-------  1 root root    0 Jul 21  2014 principal~.ok

In particular the size of the file principal~ is always '508K'.
( My master is a x86 machine.
  I have tried and a x86 slave and a x86_64 one)


<<Good news>>
If I make the file kdb-dump smaller (basically I consider a smaller
numeber of principals), the command "kdb5_util load kdb-dump" does succeed.

root at tt-u1404:/var/lib/krb5kdc# ls -larth
...
-rw-------  1 root root 8.0K Jul 22 16:33 principal.kadm5
-rw-------  1 root root    0 Jul 22 16:33 principal.kadm5.lock
-rw-------  1 root root  20K Jul 22 16:33 principal
drwx------  7 root root 4.0K Jul 22 16:33 .
-rw-------  1 root root    0 Jul 22 16:33 principal.ok

And I can
- start the daemons krb-kdc and krb-admin-server
- get tickets and ssh to client-machines
(where
 kdc = tt-u1404.doc.ic.ac.uk)

Does it ring any bell?

Cheers,
Giuseppe



More information about the Kerberos mailing list