[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Questions regarding database replications


I am working on a test cell migrating from an afs kaserver to heimdal
krb5 version 0.6.1 on solaris 9.

It seems that using iprop breaks the kaserver emulation.  I am able to
sucessfully propogate the the database to the slave servers using iprop.
When I use kinit I get a working ticket and afs token as expected.  If I
try to klog I get errors that the afs user doesn't exist in the
database. I am running the kdc on the AFS database servers and kaserver
is shutdown.  I run the kdc and iprop from the bosserver.

This problem does not occur when I use hprop to propogate the database
to the slave servers.

I have some other questions that I have note been able to find answers
to in the docs -

In Jason Garman's book on kerberos he writes that the database file
is /var/heimdal/heimdal.db.  I don't see such a file.  The following is
a listing of /var/heimdal on my master server --

-rw-r--r--   1 root     other      25432 Apr 20 13:30 default.log
-rw-------   1 root     other       4096 Apr 20 11:58 heimdal.dir
-rw-------   1 root     other          0 Apr 20 11:11 heimdal.lock
-rw-------   1 root     other    28046336 Apr 20 13:30 heimdal.pag
-rw-------   1 root     other          0 Apr 20 11:15 heimdal~.lock
-rw-r--r--   1 root     other         18 Apr 20 11:19 kadmind.acl
-rw-r--r--   1 root     root       24539 Apr 20 13:40 kdc.log
-rw-------   1 root     other      49039 Apr 20 13:30 log
-rw-------   1 root     other         69 Apr 20 11:10 m-key
srwxr-xr-x   1 root     root           0 Apr 20 11:27 signal
-rw-r--r--   1 root     other         69 Apr 20 11:29 slaves
-rw-r--r--   1 root     root         371 Apr 20 13:43 slaves-stats

I assume the actual database is heimdal.pag.  Is this correct??

I am somewhat suprised to find that the sizes of heimdal.pag are
different on the slaves.

-rw-------   1 root     root      800768 Apr 20 13:30 heimdal.pag

The size also changes when I propgate the database using hprop

-rw-------   1 root     root      515072 Apr 20 13:57 heimdal.pag

I would have assumed that the database would be exactly the same on the
master and slaves regardless of the method of propagation.

Gedaliah Wolosh, Ph.D.
Manager Computing Resources - CCS
New Jersey Institute of Technology	 Office 973 596-5437
323 King Blvd 	GITC 2203		 Fax    973 642-4761
Newark, NJ 07102                         Email  gwolosh@njit.edu