Commvault Unveils Clumio Backtrack - Near Instant Dataset Recovery in S3

// 18 Sep 2020

Don’t Risk RDS Data Loss: Diversify Backups

Abdul Rasheed
ShareTwitterfacebookLinkedin

Time and time again it happens for organizations using Amazon RDS. No, the problem is not with RDS itself, which is an excellent database-as-a-service offering from AWS. The problem is with the way RDS data is getting backed up, and those backups becoming useless when your enterprise requires it.

It happened again, this time at on a massive scale. Last week, several versions of Amazon Aurora PostgreSQL disappeared from the AWS RDS portal for days. This is equivalent to engines in your cars disappearing. Your car is still there, but you can’t start it anymore. This resulted in disruptions and it took considerable time and effort for customers to work around the problem while AWS was diligently working on fixing the problem.

This is where backups should have helped, but only if you had a true RDS data backup as we deliver in Clumio. The backups shouldn’t depend on your original engine or engine version. Thus, Clumio’s customers were the only lucky ones last week, they could simply retrieve RDS data from backups and load it onto surviving database engine versions, and their applications could go back into business. How about customers without Clumio for RDS?

It is sad to see that pretty much all backup software vendors have taken a shortcut approach to protecting RDS databases. Instead of liberating backup data from backend platforms (in this case, Amazon RDS and Amazon Aurora engines), they simply perform snapshots of RDS databases where the snapshots themselves are dependent on the original engines for recovery. So when the engines disappear from the mighty AWS cloud, your so-called backups (which are really snapshots) become useless when you need it the most.

As I mentioned before, this is not the first time that customers fell into engine dependent snapshots and found themselves lost at sea. This happens all the time because AWS does need to retire older engine versions periodically. It happened last June and we wrote a story about a Clumio customer who survived that uncertainty.

So please, when you adopt RDS (and you should, it is a magic potion that makes your applications healthy without you performing the heavy lifting of nurturing a database engine), don’t ignore your data backup strategy. Snapshots are not backups. Today, Clumio is the only backup company that has thought this through and solved it right. As a matter of fact, we do use Amazon Aurora PostgreSQL in our own SaaS platform, and we are protecting it using our own service. This is the way we are ensuring data portability and independence from engine types and versions while taking advantage of the benefits of RDS databases.

Stay safe, stay SaaSy.