University of California at Berkeley
           Department of Electrical Engineering & Computer Sciences
                     Instructional & Electronics Support Group


/share/b/pub/nisplus.help
                                                                July 1, 2004

FAQ on some common problems and questions concerning NIS+

Fixing moebius from a wedged state
----------------------------------

Symptom: Moebius or any NIS+ server fills it's /var directory because the
NIS+ transactions logs have grown so big as to prevent the NIS+ server
from updating. This causes the machines which bind to it to receive old
information with regards to passwds or user information.

Fix: Stop the NIS+ services cleanout the log files from /var/nis/data
and restart the NIS+ services. After which you re-sync the replica with
the master.

Here is the commands you issue to do this...

  % /etc/init.d/rpc stop		;; Stops the nis+ services
  % cd /var/nis/data; rm -rf *.log	;; Remove the logs
  % /etc/init.d/rpc start		;; Start the nis+ services
  % nisping -C -a			;; RE-sync the replica with the master

Correcting re-sync errors
-------------------------

Symptom: you change your password on po, but other machines don't notice

Possible Cause: moebius's disk is full, and nisd will refuse to resync
with other servers when this happens.

Solution: Something you can try if moebius's partitions start to fill
up and the moebius NIS server doesn't resync itself with the Cory Hall
servers anymore is:

 1. Stop the NIS servers using /etc/init.d/rpc stop
 2. Remove all the /var/nis/data/*.log files
 3. Restart the NIS servers using /etc/init.d/rpc start
 4. Re-run the checkpoint procedure using nisping -C -a

More to come later......