Hello,
would like to start discussion about companies practises with Vertica disaster recovery plans.
decided not to use native Vertica backup restore vbr.py utility because need identical topology when restoring database. Its expensive to have a separate identical cluster for restores only.
I think about those 3 options for disaster recovery solutions:
1. Dual load - Active/Active approach. No need to have identical clusters. Need additional effort for ETL and releases. Is practical for 24X7 ie when one cluster is down for maintenance, upgrading or so on. I hope extra licence for dual load is not needed?
2. Develop a script and periodically export database objects (or object partitions) to txt/gzip files. No need additional licence. Unable to run "incremental" backups. If object DDL changes need to rerun backup.
3. Develop a script and copy periodically database objects (or object partitions) to separate Vertica cluster. Using "EXPORT TO VERTICA" using filter clause to select only latest partitions. Unable to run "incremental" backups. Same question about licencing.
Does anybody uses any of these approach or can recomend smth else?
Thanks,
Ramunas
Vertica disaster recovery plan options
+2With regards to licensing on all your points, you don't need an additional license provided that you're using the secondary cluster for DR. There are some minor caveats to that, but you should speak to your technical representative.
Dual load is the most common DR strategy. In this strategy, you have a primary cluster, and a DR cluster which are ingesting the same data.
Perhaps you're not aware, but you do not need an exact topology for the native backup/restore if you're doing object level backups.
The more important question you should be asking is "What is my SLA?" Basically, how quickly do you need to recover in the event of a disaster. The other factor is how much data loss (if any) is acceptable?
The documentation covers these strategies at a high level.
Dual load is the most common DR strategy. In this strategy, you have a primary cluster, and a DR cluster which are ingesting the same data.
Perhaps you're not aware, but you do not need an exact topology for the native backup/restore if you're doing object level backups.
The more important question you should be asking is "What is my SLA?" Basically, how quickly do you need to recover in the event of a disaster. The other factor is how much data loss (if any) is acceptable?
The documentation covers these strategies at a high level.
Sign up
Already have an account? Login
Welcome to the Rocket Forum!
Please log in or register:
Employee Login | Registration Member Login | RegistrationEnter your E-mail address. We'll send you an e-mail with instructions to reset your password.
77 4th Avenue
Waltham, MA 02451 USA
