Skip to content

Aerospike Backup Control release notes

For the complete documentation index see: llms.txt

All documentation pages available in markdown.

absctl 1.1.0

July 16, 2026  |  Download

Summary

This release is built on backup-go v0.11.0. It adds file logging and tighter configuration validation. absctl also reports finer-grained progress during long-running backups and restores.

New features

  • Add --log-file to write logs to a file instead of stderr. Use it with --log-json to write JSON log lines. absctl appends to the file and does not rotate it, so configure external rotation and retention for long-running jobs. [BKRS-46]

Improvements

  • Disallow using --continue together with --remove-files, because removing files invalidates the saved backup state. [BKRS-71]
  • Validate --max-records and reject the flag unless --parallel is exactly 1. [BKRS-10]
  • Validate that --unique and --replace are mutually exclusive on restore. [BKRS-105]
  • Validate cloud backup --file-limit against chunk size so multipart uploads stay below 10,000 parts. [BKRS-187]
  • Validate encryption configuration more strictly: require --encrypt with a key source and allow only one key source. [BKRS-129]
  • Retrieve additional Aerospike Database connection and cloud-storage parameters from the Aerospike Secret Agent. [BKRS-89]
  • Retrieve --tls-* cluster connection flags from the Aerospike Secret Agent. [BKRS-129]
  • Rename structured log attributes to kebab-case, including parallel-read, file-limit, and state-file-dst. Update log parsers that relied on the previous names. [BKRS-74]
  • Show backup and restore progress with two decimal places, such as 99.31%, and report it more frequently during long jobs. [BKRS-218]
  • Increase default cloud upload chunk sizes to 50 MiB for Amazon S3, Google Cloud Storage (GCS), and Azure Blob Storage. Change the default local file buffer to 4 MiB. [BKRS-187]
  • Improve scan throttling to limit backup load on the cluster. [BKRS-8]
  • Retry transient failures when creating secondary indexes and user-defined functions (UDFs) during restore. [BKRS-145]
  • Print backup estimates consistently and show the number of files before a restore starts. [BKRS-190]
  • Optimize Amazon S3 backup memory usage. [BKRS-155]
  • Optimize restore decoder performance. [BKRS-182]
  • Reduce allocations when encoding backup tokens.
  • Render duration values in logs as human-readable strings, such as 30s and 1m30s, instead of raw nanoseconds. [BKRS-171]
  • Write service messages to stderr so stdout stays clean for piped output.
  • Remove the misleading chunk-size log message during restore. [BKRS-231]

Bug fixes

  • Fix authentication to Google Cloud Storage when running with an attached Google Cloud service account and no key file. [BKRS-40]
  • Retry transient errors during large Google Cloud Storage uploads. [BKRS-244]
  • Fix a deadlock that could occur during secondary index creation on restore by updating the Aerospike Go client. [BKRS-26]
  • Improve encryption key error messages when PEM decoding fails. [BKRS-129]
  • Release limiter slots when a backup is canceled so later backups can start. [BKRS-114]
  • Change record-count mismatch messages from error to warning severity. [BKRS-156]
  • Remove internal client IDs from log messages. [BKRS-148]

For the complete changelog, see the absctl 1.1.0 GitHub release page.

absctl 1.0.0

February 2, 2026  |  Download

This initial release introduces absctl, the new Aerospike Backup Control command-line tool. Designed to replace legacy asbackup and asrestore, absctl consolidates backup and restore workflows into a single utility featuring native cloud support and improved operational efficiency.

Highlights

Unified CLI

  • Backup and restore consolidated into one tool: absctl backup and absctl restore
  • YAML configuration file support for scripted and scheduled operations
  • Structured JSON logging for log aggregation pipelines

Improved partition-based scanning

Backups use balanced partition-based scans with configurable parallelism (--parallel) for efficient hardware utilization, ensuring minimal and balanced load across all Aerospike cluster nodes:

  • Parallel partition scanning: Configurable parallelism (1-32 workers) with intelligent partition distribution
  • Partition filtering: Back up specific partition ranges or individual partitions
  • Minimal cluster impact: Rate limiting (RPS and bandwidth) to prevent backup operations from affecting production workloads
  • Resumable backups: State file support for continuing interrupted backups from the last checkpoint

Storage backend support

Flexible storage options for backup and restore:

  • Local disk: Direct backup to local filesystem with configurable buffer sizes
  • Amazon S3: Full support including storage classes (Standard, Glacier, Intelligent-Tiering, etc.), multipart uploads, and MinIO compatibility via endpoint override
  • Google Cloud Storage (GCS): Native GCP bucket integration
  • Azure Blob Storage: Complete Azure container support with both account key and Azure Active Directory authentication

Cloud backups do not require intermediate local storage.

AZ-aware backup

Backup node selection to optimize network topology and reduce costs:

  • --prefer-racks: Prefer reading from specific rack IDs while backing up the entire namespace (can fall back to other racks)
  • --rack-list: Back up partitions whose master is on a node in the listed racks when the job starts (potentially partial; restore all slices for a full data set)

--prefer-racks can reduce cross-availability-zone traffic during a full backup. --rack-list supports parallel backup sliced by rack. See Rack-aware backup: prefer-racks vs rack-list.

Additional features

  • TLS/mTLS for cluster connections
  • AES-128/256 encryption for backup files
  • ZSTD compression
  • Aerospike Secret Agent integration for credential management

Backward compatibility

  • Backup files (.asb) created by asbackup can be restored with absctl
  • Most CLI flags preserved from legacy tools

For migration details, see Legacy migration.