Skip to content

Releases: ceph/ceph-csi

Ceph CSI v2.1.0 Release

15 Apr 13:20
Compare
Choose a tag to compare

Release Issue #806

Changelog or Highlights:

Features:

Add support for rbd static PVC
Move cephfs subvolume support from `Alpha` to `Beta`.
Added support for rbd topology-based provisioning.
Support externally managed configmap.
Updated Base image to ceph Octopus
Added csiImageKey to keep track of the image name in RADOS omap
Added E2E for helm charts
...

Enhancements:

Implement CreateVolume with go-ceph which boosts performance.
Migrated from `dep` to `go modules`
Updated Kubernetes version to v1.18.0
Updated golang version to 1.13.9
Updated Kubernetes sidecar containers to latest version
E2E: Add Ability to test with non root user
...

Bug Fixes:

Log an error message if cephfs mounter fails during init
Aligned with klog standards for logging
Added support in to run E2E in a different namespace
Removed cache functionality for cephfs plugin restart
rbd: fallback to inline image deletion if adding it as a task fails
code cleanup for static errors and unwanted code blocks
Fix mountoption issue in  rbd
travis: re-enable running on arm64
...

Deprecated:

GRPC metrics in cephcsi

Documentation:

Added Document to cleanup stale resources
Updated ceph-csi support matrix
dev-guide: add reference to required go-ceph dependencies
Update upgrade doc  for node hang issue
....

Ceph CSI v2.0.1 Release

19 Mar 17:16
Compare
Choose a tag to compare

Release Issue #855

Changelog or highlights:

  • Fix rbd mount option issue when mount options from storage-class are not passed to mount point
  • Added a check for cephfs subvolume present in the backend when creating a PVC
  • Added forcecephkernelclient as startup parameter in helm chart
  • Fix PVC delete issue when the pool is deleted
  • Fix typo in pod spec in rbd helm chart
  • Add _netdev as default mount options for both cephfs and rbd
  • Check for ENOENT errors when deleting CephFS volumes
  • Fix ceph-csi image tag version in rbd helm chart
  • Added PodSecurityPolicy support to helm charts

Ceph CSI v2.0.0 Release

20 Jan 14:00
Compare
Choose a tag to compare

Release Issue #557

  • Added dynamic resize support for CephFS PVCs
  • Added dynamic resize support for RBD PVCs
  • Added encryption with LUKS support for RBD PVCs
  • Mutli arch support ( ceph-csi Arm64 image)
  • Upgrade documentation from v1.2.2 to v2.0.0
  • Updated code base to kube v1.17
  • leader election enabled in deployment
  • Added Version flag to cephcsi
  • Removed Kubernetes 1.13.x support with 2.0.0 release
  • CSI: run all containers as privileged in daemonset pods
  • Upgrade: csi-attacher sidecar from v1.2.0 to v2.1.0
  • Upgrade: csi-snapshotter sidecar from v1.2.1 to v1.2.2
  • Upgrade: csi-node-driver-registrar sidecar from v1.1.0 to v1.2.0
  • Upgrade: csi-resizer from sidecar v0.3.0 to v0.4.0
  • Update csi-provisioner from sidecar v1.3.0 to v1.4.0
  • Remove deprecated containerized flag in rbd
  • Discard umount error if the directory is not mounted
  • Use EmptyDir to store provisioner socket
  • Add ContentSource to the CreateVolume response
  • Rbd: only load nbd module if not available yet
  • Enhance scripts to deploy ceph cluster using rook
  • Add e2e tests for RBD resizer
  • Update minikube to latest released version
  • Update golangci-lint version to v1.21.0
  • Fix to use kubectl create not kubectl apply in the e2e
  • Add volume size roundoff for expand request
  • Add E2E for cephfs resize functionality
  • Add Documentation for PVC resize
  • Fix block resize issue in RBD
  • Add 13.0.0 Mimic supported version to the readme
  • update Metrics supported version in Readme
  • Remove hard-coded UpdateStrategy from templates
  • Add E2E for block PVC resize
  • Enable logging in E2E if the test fails
  • Enable Block E2E for rbd
  • Add ID-based logging for ExpandVolume
  • Validate rbd image name in NodeExpand

Ceph CSI v1.2.2 Release

21 Nov 05:03
Compare
Choose a tag to compare

Release Issue #639

Changelog or highlights:

  • Add kernel version detection for mounting with client
  • Fix issue in helm chat pushing
  • Added forcecephkernelclient as startup parameter to force enabling ceph
  • Handle EACCESS error from 'ceph rbd task add remove'
  • Fix vol size for cephfs and rbd
  • Update registration directory name
  • Reorder kernel version checking logic
  • Increase sidecar time out value to 150s.
  • Format ext4 with lazy_journal_init
  • Change default csi liveness ports to ones less common
  • Add POD_NAMESPACE to RBD provisioner deployments
  • Add a check for nil secrets
  • Remove rootfs from rbd provisioner pod
  • Added E2E for rbd nodeplugin restart
  • Deprecate containerized flag and remove nsenter

CSI v1.2.1 Release.

25 Sep 10:12
Compare
Choose a tag to compare

Release issue # #600

*) Change the recommended/default FS for RBD to ext4
*) Use nodiscard option while formatting RBD devices.
*) Use provisioner socket while probing liveness.
*) Reject request if the operation is in progress
*) Fix pod termination issue due to stale mount after node plugin restart.
....etc

Ceph CSI v1.2.0 Release

06 Sep 17:17
c420ee6
Compare
Choose a tag to compare

Release Issue # #393

Changelog or highlights:

*) Cephfs: Use ceph kernel client if kernel version >= 4.17
*) implement grpc metrics for ceph-csi
*) Add xfs fstype as default type in storageclass
*) Add support to use ceph manager rbd command to delete an image
*) e2e: correct log format in execCommandInPod()
*) Add 'gosec' to the static-checks
*) switch to cephfs, utils, and csicommon to new loging system
*) utility to trace backend volume from RBD pvc
*) Implement context-based logging
*) implement klog wrapper
*) unmap rbd image if connection timeout.
*) start controller or node server based on config
*) fix: Adds liveness sidecar to v1.14+ helm charts
*) Prometheus liveness probe sidecar
*) Wrap error if failed to fetch mon
*) provisioners: add reconfiguring of PID limit
*) Use "rbd device list" to list and find rbd images and their device paths
*) Update Unstage transaction to undo steps done in Stage
*) Move mounting staging instance to a sub-path within staging path
*) e2e: do not fail to delete resources when "resource not found"
*) remove post validation of rbd device

Many other bug fixes, code improvements, README updates are also part of this release.

Ceph CSI v1.1.0 Release

12 Jul 08:34
c7ba26d
Compare
Choose a tag to compare

Changelog or highlights:

*) CephFS subvolume/manager based volume provisioning and deletion.
*) E2E test support for PVC creation, App pod mounting.etc.
*) CSI spec v1.1 support
*) Added support for kube v1.15.
*) Configuration store change from configmap to rados omap
*) Mount options support for CephFS and RBD volumes
*) Move locks to more granular locking than CPU count based
*) Rbd support for ReadWriteMany PVCs for block mode
*) Unary plugin code for 'CephFS and RBD' drivers.
*) Driver name updated to CSI spec standard.
*) helm chart updates.
*) sidecar updates to latest available.
*) RBAC corrections and aggregated role addition.
*) Lock protection for create,delete volume ..etc operations
*) Added external snapshottter to CSI.
*) Added CSIDriver CRD.
*) Support matrix table addition in documentation.
*) Many linter fixes and error code fixes.
*) Removal of dead code paths.
*) StripSecretInArgs in pkg/uti
*) MIgration to klog from glog.

Many other bug fixes, code improvements, README updates are also part of this release.