Skip to content
Maintained by AxonOps — production-grade documentation from engineers who operate distributed databases at scale Simplify Cassandra Ops Simplify Kafka Ops

Overview

AxonOps provides the ability to restore from local snapshots and remote backups.

The Restore feature is accessible via Operations > Restore.

restore

Note that the AxonOps user needs temporary write access to Cassandra data folders to perform the restore.

To restore Cassandra, click the backup you want to restore.

This provides the backup details and lets you start the restoration by clicking LOCAL RESTORE or REMOTE RESTORE, depending on whether you want to restore from the local snapshot or the remote backup (if remote backups are configured). Here you can also select a subset of nodes to restore via the checkboxes in the Nodes list.

restore

Multiple data directories

Requires axon-agent 2.0.23 or later.

If your Cassandra nodes use more than one entry under data_file_directories in cassandra.yaml, restores work as follows:

  • Local restore — SSTables are restored to the same data directory they were snapshotted from, preserving the original on-disk layout.
  • Remote restore — SSTables are placed across the configured data directories using a round-robin policy, balancing the load between disks. All component files belonging to a single SSTable generation (for example nb-12-big-Data.db, nb-12-big-Index.db, nb-12-big-Statistics.db) are kept together on the same data directory so Cassandra can read them as a unit.

The AxonOps user needs temporary write access to every directory listed in data_file_directories. See the per-scenario restore guides for the exact chmod/chown commands.

Follow the links below for detailed restore scenarios:

Restore a single node - same IP address

Replace a node - different IP address

Restore whole cluster - same IP addresses

Restore whole cluster - different IP addresses