2

I am working with a Hyper-Converged Database (HCD), which is based on Cassandra, and I need to implement a backup strategy that includes commitlogs along with SSTables. I understand that Medusa only captures SSTables, and without commitlogs, I cannot perform point-in-time recovery. My requirement is to back up commitlogs in addition to SSTables so that I can restore the database to its exact state before a failure. Specifically, I want to know:

  1. What are the recommended ways to back up commitlogs in HCD (or Cassandra)?

  2. Can tools like Medusa handle commitlog archiving automatically?

  3. Alternatively, can I configure cassandra.yaml for commitlog_archiving to copy logs to an external storage location?

I’ve reviewed the documentation, but could not find a detailed guide on commitlog backup for HCD.

So far, I have: Taken snapshots using Medusa for SSTables. Checked Medusa’s documentation for commitlog backup, but need clarification on the configuration. Looked at the commitlog_archiving settings in cassandra.yaml, but not sure about the correct procedure for HCD.

Expected result: A reliable method to automatically back up commitlogs along with snapshots (using Medusa or native configuration).

Actual result: I only see SSTables backed up; commitlogs are not included in the current backup process.

Improve this question

1 Answer 1

Reset to default
1

You’re right that Medusa only handles SSTables and doesn’t archive commitlogs, so it can’t provide point‑in‑time restore on its own. The standard way to achieve PITR in Cassandra is to combine snapshots with commitlog archiving (via the built‑in commitlog_archiving settings), but that’s something you’d need to wire up and maintain yourself alongside Medusa.

Since you mentioned this is important for your setup, you might also want to take a look at AxonOps.

Disclosure: I work there.

AxonOps includes continuous commitlog archiving and full point‑in‑time recovery out of the box as well as extensive monitoring and alerting plus integrations for teams, slack, snow, PagerDuty etc

We haven’t tested it specifically with HCD, but if you’re running Apache Cassandra directly, AxonOps can manage all your backups and operations. Just thought I’d share in case PITR is critical in your environment.

Improve this answer

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.