Skip to content

CHANGELOG.md

Latest commit

 

History

History
157 lines (124 loc) · 5.78 KB

CHANGELOG.md

File metadata and controls

157 lines (124 loc) · 5.78 KB

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[Unreleased]

[1.2.8]

Added

  • OpenTelemetry Tracing: Integrated OpenTelemetry-based tracing across all API calls using standard W3C traceparent propagation.

  • Injected trace ID into all HTTP headers for REST API communication with Hammerspace.

  • Configured otel.TracerProvider using the hammerspace-csi instrumentation scope for enhanced observability and trace correlation.

  • NFS Root Share Mounting: Introduced support for mounting a root share for NFS-backed volumes.

  • Volumes now use bind mounts from this root instead of individual share mounts per volume.

  • mountBackingShareName Parameter: Added support for mountBackingShareName in the NFS storage class.

  • When specified, the actual NFS volume directories will be created inside this parent share.

  • Improved Logging: Added more detailed log messages throughout the CSI plugin for better operational clarity and troubleshooting.

Changed

  • Mount Behavior for Volume Types:

    • For NFS volumes, the plugin mounts a single root share once per node and uses bind mounts per volume.

    • For file-backed and block volumes, the CSI driver continues to mount the share defined in the storage class and then perform bind mounts as before.

Fixed

  • Ensured trace ID is not dropped during API retries or chained calls.

Security

  • Reviewed HTTP client configuration for trace propagation compatibility with secure endpoints.

1.2.7

Fixed

  • Resolved an issue in NodeGetVolumeStats where excessive backend GetShare API calls were triggered for NFS volumes, causing SM log flooding. The function now uses syscall.Statfs directly on the volume mount path to obtain usage metrics, reducing API load.
  • Improved CleanupLoopDevice to retry loop device detachment up to 3 times with a 1-second interval, ensuring better reliability when devices are temporarily busy.

Added

  • Introduced support for configurable unmount retry behavior using environment variables (UNMOUNT_RETRY_COUNT, UNMOUNT_RETRY_INTERVAL), which can be injected via Kubernetes ConfigMap.
  • Enhanced FIP (Floating IP) selection logic to support strict round-robin ordering for multi-portal NFS mounts:
    • The CSI driver now maintains a rotating index to evenly distribute data access across available portal IPs.
    • This reduces hotspotting and improves throughput in clusters with multiple floating IPs.
    • If an FQDN is configured and resolves to a reachable NFS endpoint, it is used directly; otherwise, the round-robin FIP selection is attempted in order.

Changed

  • Production image: Replaced CentOS 8 UBI base image with Rocky Linux 9 UBI for better long-term support and compatibility with modern Python 3 and security patches.
  • Development image: Updated Dockerfile_dev to use golang:1.24-alpine, removed CentOS dependencies, and transitioned to musl-based Alpine packages. Python hstk tool is now installed in a virtual environment to comply with PEP 668.

1.2.6

Fixed Bug

  • Fixed error where floating IP's is not being used.

1.2.5

Added

  • Fixed error to get volume capability due to change in type fromat. (Fix breaking changes only work with thor2 and above)
  • Added List volumes support
  • Tested working with go 1.21.3 and k8 v1.27.4

1.2.4

Added

  • Fixed error while creating share to track task status.
  • Added check for nfs mount before attempting to delete nfs mount volume.
  • Update docker dev file to use go lang v:1.21 and gocsi v:1.2.2

1.2.3

Added

  • Updated the deprecated module
  • Added url parse to URI string, it was crashing the CSI when share name have % in it
  • Added condition to expand volume only when share state is Mounted state

1.2.2

Added

  • Added share name length restriction to 80 characters.

1.2.1

Added

  • Removed unnecessary mount option

1.2.0

Added

  • Supports online resize of file-backed devices
  • Switch to UBI image
  • Version update to UBI 8.4

1.0.4

Added

  • Support for Portal Floating IPs

1.0.3

Added

  • Support for share descriptions

1.0.2

Added

  • Support for Hammerspace 4.4.0
  • Tested with Kubernetes up until 1.18
  • Support for resize for NFS (without remount) and file-backed (requires remount) volumes

1.0.1

Added

  • Support for Hammerspace 4.3.0
  • Tested with Kubernetes 1.16, 1.17

1.0.0

Added

  • Topology key topology.csi.hammerspace.com/is-data-portal
  • Ability to set additional metadata tags on plugin created shares and files
  • Kubernetes 1.14 deployment manifests

Changed

  • Golang version 1.10 -> 1.12

0.1.3

Added

  • Default size (1GB) for file-backed volumes
  • Support for filesystems other than nfs for Mount Volumes

Changed

  • Wait for file-backed volumes to exist on metadata server before responded successful for create
  • Return error if source snapshot does not exist

Fixed

  • Issue where block volume snapshots may not be properly deleted

0.1.2

Added

  • Ability to specify export path prefix to use when mounting to a data portal HS_DATA_PORTAL_MOUNT_PREFIX
  • Command execution timeout of 5 minutes
  • Support for CSI spec v0.3 (Kubernetes 1.10-1.12)

Changed

  • Combined Kubernetes Deployment yaml files
  • HS_TLS_VERIFY defaults to false
  • Require HTTPS communication with Hammerspace API

Fixed

  • Set objectives on file for block volumes

0.1.1

Fixed

  • Include required CA root certificates in docker image

0.1.0

Added

  • Support for CSI spec 1.0
    • Mounted Volumes
      • Create
      • Delete
      • Snapshot
      • Publish/Unpublish
    • Block Volumes
      • Create
      • Create from snapshot source
      • Delete
      • Snapshot
      • Publish/Unpublish