Backup and recovery is one of the primary objectives of this group. We have designed a general plan that applies to all supported databases, regardless of DBMS or operating system. Database backups are taken at regular intervals (see specifics below). In most cases, copies of these regular backups are kept on-line (disk) for up to two weeks. Additionally, when possible backup files are also stored on tape for as many as 30 days at a time. We maintain 2 sets of tapes for our backups: one set in the CIS machine room/tape silo and another set off-site for Disaster Recovery purposes.
For production SQL Servers, full backups of each database are taken nightly. In addition, the transaction logs are backed up hourly. For test or development SQL Servers, full backups of each database are taken weekly while transaction logs are backed up nightly.
For the ADABAS DBMS, full backups are taken nightly. Protection Logs (p-logs) are written to disk as they fill up.
For the MySQL RDBMS, both full database backups & binary log backkups are taken nightly. Dual copies of these backups are sent directly to the z/Os tape system.
For the Oracle RDBMS, full backups are taken nightly of all production databases. Where possible, RMAN is used as the primary means of backup control. Where RMAN is not in operation, we use a variety of backup methods including cold file sytem backups and NetApp Snapshots.
It is considered good practice to perform an ad-hoc backup before any special processing or before significant changes are made to a database. This provides a good starting point for any recovery operations that may be needed.
With at least 48 hours prior notice, the database team can schedule an automated backup to run prior to these events. Please use our Database Team Request Form to notify the appropriate DBA of your ad-hoc backup needs.
Should a situation arise that requires a backup with out advance notice, please contact Help Desk Central and have them contact your DBA directly.