Release Note
DON'T UPGRADE from/to any RC/Preview releases, because the operation is not supported.
v1.4.4-rc2 released! 🎆
This release introduces bug fixes and improvements, mainly related to stability. Please try it and feedback. Thanks for all the contributions!
Installation
Please ensure your Kubernetes cluster is at least v1.21 before installing v1.4.4.
Longhorn supports three installation ways including Rancher App Marketplace, Kubectl, and Helm. Follow the installation instructions here.
Upgrade
Please read the important notes first and ensure your Kubernetes cluster is at least v1.21 before upgrading to Longhorn v1.4.4 from v1.3.x/v1.4.x, which are only supported source versions.
The live upgrade from v1.3.x/v1.4.x is not supported until the GA release. When available, follow the upgrade instructions here. Here.
Deprecation & Incompatibilities
N/A
Known Issues after Release
Please follow up on here about any outstanding issues found after this release.
Resolved Issues
Enhancement
- [FEATURE] Add disk status prometheus metrics 6858 - @c3y1huang @chriscchien
Improvement
- [IMPROVEMENT] Old kernel such as 3.10.0 set provisioning_mode to wrong value (writesame_16, disabled, full, ...) but not the correct value (unmap) so the trim feature doesn't work 6854 - @PhanLe1010 @chriscchien
- [IMPROVEMENT] Improve log level for resource update failure able to reconcile again 6843 - @PhanLe1010 @nitendra-suse
- [IMPROVEMENT] Prevent unexpected engine creation 6682 - @PhanLe1010 @ejweber @roger-ryao
- [IMPROVEMENT] Support both NFS
hardandsoftwith customtimeoandretransoptions for RWX volumes 6655 - @derekbit @roger-ryao - [IMPROVEMENT] Prevent Volume Provision if Related Backing Image Stuck in Ready-For-Trasfer State 6615 - @ChanYiLin @roger-ryao
- [IMPROVEMENT] Remove dummy services of each CSI sidecar if not required 6581 - @ejweber @roger-ryao
- [IMPROVEMENT] Include /var/log/messages during the support-bundle syslog collection 6544 - @c3y1huang @roger-ryao
- [IMPROVEMENT] Provide more information for volume scheduling failure 6461 - @smallteeths @chriscchien
- [IMPROVEMENT] Improve upgrade path and make it more solid 6294 - @PhanLe1010 @roger-ryao
- [IMPROVEMENT] Fix scheduling flooding logs 6019 - @ChanYiLin @roger-ryao
- [IMPROVEMENT] Unify logs with extra static info like module/method/function/line 5509 - @ChanYiLin @roger-ryao
- [IMPROVEMENT] Add pvc name to longhorn_volume metrics 5297 - @c3y1huang @nitendra-suse
- [IMPROVEMENT] Avoid the accident deletion of longhorn settings 4984 - @ejweber @roger-ryao
- [IMPROVEMENT] Remove Longhorn engine path mismatch log 3786 - @c3y1huang @roger-ryao
Stability
- [BUG] DR volume failed when synchronizing the incremental backup 6750 - @mantissahz @chriscchien
- [BUG] Somehow the Rebuilding field inside volume.meta is set to true when one replica only, causing the volume into attaching/detaching loop 6626 - @c3y1huang @nitendra-suse
Resilience
- [BUG][v1.4.4-rc1] rwx workload gets stuck in ContainerCreating after cluster restart 6924 - @yangchiu @derekbit
- [BUG] Volumes failing to mount because of engine upgradedReplicaAddressMap reference 6762 - @PhanLe1010 @chriscchien
- [BUG] Share manager pod will stay in IO error when the volume becomes read only 5961 - @ChanYiLin @roger-ryao
Bug
- [BUG] IO error occurs when detaching RWX volume 6829 - @derekbit @chriscchien
- [BUG] Two active engine when volume migrating 6642 - @PhanLe1010 @chriscchien
- [BUG] Longhorn Instance Manager Memory leak 6481 - @james-munson @chriscchien
- [BUG] The instance manager with state unknown will be cleaned up in the split-brain case 6479 - @shuo-wu
- [BUG] Fix errors in questions.yaml 6392 - @james-munson @chriscchien
- [BUG] Webhook is never called for BackingImageManager 6328 - @ejweber @chriscchien
- [BUG] Environment Check Script Fails To Perform All Checks 5653 - @PhanLe1010 @roger-ryao
- [BUG] Can't delete volumesnapshot if backup target not set 4979 - @ejweber @chriscchien
- [BUG] Backup Job returns "Completed" despite running into errors 4255 - @mantissahz @chriscchien
- [BUG] Error during backup process will be removed quickly without user knowing 1249 - @mantissahz @chriscchien
- [BUG] longhorn manager isn't annotated with iam.amazonaws.com/role 6947 - @mantissahz
- [BUG] [v1.4.4-rc1] Wrong image name in uninstall manifest(uninstall job) 6895 - @innobead @chriscchien
Misc
- [TASK] Revert "Disable Automatically Delete Workload Pod when The Volume Is Detached Unexpectedly for RWX volumes" 6838 - @derekbit @roger-ryao