Veritas NetBackup™ Vault Administrator's Guide
- About Vault
- Installing Vault
- Best Practices
- About preferred vaulting strategies
- About how to ensure that data is vaulted
- About not Vaulting more than necessary
- About preparing for efficient recovery
- About avoiding resource contention during duplication
- About how to avoid sending duplicates over the network
- About increasing duplication throughput
- About organizing reports
- Configuring NetBackup Vault
- Configuring Vault
- About Vault configuration
- About configuring Vault Management Properties
- About creating a vault
- About creating profiles
- Configuring a profile
- Vaulting and managing media
- About Vault sessions
- About monitoring a Vault session
- About the list of images to be vaulted
- About ejecting media
- About injecting media
- About using containers
- About vaulting additional volumes
- About using notify scripts
- Creating originals or copies concurrently
- Reporting
- Administering Vault
- About administering access to Vault
- About NetBackup Vault session files
- Using the menu user interface
- Troubleshooting
- Debug logs
- Appendix A. Recovering from disasters
- Appendix B. Vault file and directory structure
About overlapping the time window in the profile
To ensure that all data is vaulted, overlap the time window in the profile.
A Vault profile uses a time range as one of the criteria for choosing the backup images to be vaulted. Vault does not duplicate or eject a backup image that already has a copy in the off-site Volume Group; therefore, Vault does not process images that are already vaulted by a previous session. Perhaps more importantly, backups that were not processed if a previous session failed are processed when the profile runs again if the time window is long enough.
Therefore, configure the time window to be the sum of the following:
The longest expected downtime for a server or robot
Twice the length of the frequency at which the profile runs
For example, if you have a profile that duplicates images daily and your longest expected downtime is three days, configure the time window to be at least five days. If a robot fails and requires three days to repair, the next time the profile runs, it selects backup images that were not vaulted during the three-day downtime. Configuring the window to be longer, such as seven days, provides even more resiliency. A longer time window forces Vault to search a larger list of images for vault candidates. Although that consumes more processing time, the extra time may not be a problem in your environment because Vault is a batch process that does not demand immediate system response.