Scenario 4: Creating a third copy of data using backup copy job as source
Infrastructure: Medium to large infrastructure (100 or more VMs)
Policies: New tapes are loaded and used tapes are exported on a certain day of the week.
Recommended settings:
Entity | Configuration settings |
---|---|
Media pool | Full and incremental backups should use different media pools. It is possible to set longer retention for full backups media pool and shorter - for incremental backups media pool, decreasing granularity for backups (e.g. 2 months for full backup media pool, 2 weeks for incremental backup media pool) |
Media pool (retention) | Protect data for <desired period> option should be used (specify the period you need, e.g. 7 days, 1 month, etc.) |
Media set (retention) | Automatically create new media set daily on <day> (e.g. Monday, etc.) |
Source (backup job) | _Backup copy job _with copy interval of 1 day, minimum 4 points retention is required |
Tape job | Eject media upon job completion option should be enabled. Export current media set.. should be enabled for the day when used tapes are extracted from library |
Scheduling | BCJ without GFS works as forever incremental job. Tape job should run every day, Virtual full must be scheduled to run periodically, dividing the forever incremental chain in smaller "chunks" to avoid risk of overwriting the only VBK (next run the job will have copy it once again). -OR- BCJ with GFS enabled will create full backups periodically. Enable “Process latest backup chain only” in tape job options to avoid copying these points. Main chain works as forever incremental job. Schedule virtual full as suggested above -OR- BCJ with active full GFS (option “read entire point” enabled) works as forward incremental job. GFS points should be created weekly. Tape job will copy incremental and full points. Virtual full can be scheduled to run rarely as it’s effectively redundant. |
Tape job will ignore the latest point in source chain because it’s locked due to continuous nature of BCJ