Mac OS X Server 1.x: Recovering From NetInfo Disasters

Note: This article pertains to Mac OS X Server versions 1.x, which were released prior to May 2001.

1. A NetInfo database is destroyed:

The easiest thing to do is to restore the database from your backups. Be sure to do this in single-user mode with NetInfo shut down.

A backup of your NetInfo database is made every night (assuming you computer is on overnight) into /var/backups. In the event that you just did something nasty to your NetInfo configuration, you can restore yesterday's version by entering the following commands:

mv /etc/netinfo /etc/netinfo.old
cd /var/backup
pax -rw netinfo /etc

Then restart the computer.

If restoring in single-user mode is impossible, remove any vestiges of the .nidb directory you are going to restore, start up the computer, restore the database, and reboot the computer. This procedure ensures there is no netinfod trying to work with a corrupted or partial database.

2. The only clone of a domain on a given subnet becomes unavailable for a long time, perhaps because its host computer's power supply fails.

If a clone fails, the easiest thing to do is to build a new clone. See the Mac OS X Server Networking and System Administration manual for details.

3. The master NetInfo server's computer fails catastrophically.
Read "The Crash of the Master NetInfo Server"; Cottle,1993.