24. Monitoring servers

P-Synch requires very little ongoing administration. It is advisable, however, to monitor server health.

Monitoring is primarily done on the master P-Synch server. You should:

• Check service health, using the Monitor / services menu option in the nph-psa program.

• Periodically try to reset passwords for a test user that has been defined on every system. Verify that the password reset works across all systems.

• Monitor the contents of PUSHPASS.LOG in the TEMP directory of the P-Synch server. It should give few or no warnings, and no errors.

• Monitor CPU load and virtual memory consumption on the P-Synch server. Both should be stable (change only very slowly) and significantly below the total resources available on the server.

• Monitor the amount of free disk on the P-Synch server. P-Synch will fail if it runs out of disk space (unlikely on a modern server).

• Check the text of PSUPDATE.LOG on the TEMP directory of the P-Synch server. It should show that the automatic nightly process to collect account information from systems where P-Synch manages passwords was successful.

• Review recent contents of the P-Synch session log, by asking for reports on user and administrator events.

• Check the sizes of the P-Synch databases. With the exception of log databases, they should have approximately constant sizes from day to day.