^{Highley Recommended, Inc.}

System Backup Process

The engineering backup process operates as follows; each primary server neptune, starfish, urchin and moby does a complete ufsdump backup Monday through Friday. The server seahorse does an entire engineering Unix network backup Monday through Friday creating a duplicate back of the primary servers neptune, starfish, urchin and moby. On Friday we stick a second tape in the seahorse system, tape units are configured for mirror mode, which becomes the off site archive tape taken to the library to be cataloged and sent off site for storage and disaster recovery. There is a catalog and restoration instructions written to the front of each tape. To restore a file system, "cat /dev/rmt/?0u", and follow the instructions. See an example:

The back up process uses the following scripts and files: /alldump /.rhosts.bck - on seahorse /.rhosts - on each client node that seahorse backs up To add a new client system to the network wide back up edit the following files: /usr/local/etc/crontabs/HostList /.rhosts.bck - on seahorse /.rhosts - on new client To set up a new server that will have a tape unit to back up its disk drives, copy the alldump script to it and edit the "List" and "Tape" variables. List should be null to back up just the server system. The alldump script is executed from the root crontab file. Each tape unit has two drives with 125GB/tape.