Release Note
v1.1.2 released! 🎆
There are lots of bug fixes as highlighted below including stability, performance, availability, troubleshooting, and etc. Thanks for all the contributions!
- Ensure volume detachment & attachment successfully when kubelet restarts
- Before v1.1.2, when kubelet restarts, it's possible to cause volumes failed to detach and attach, then the corresponding workloads keep staying in terminating or pending state. In v1.1.2, the problem is resolved by introducing the idempotent operation to ensure volumes detach and attach successfully.
- Ensure recurring jobs run again when the volume is attached after being detached for a long time
- Before v1.1.2, the recurring job depends on CronJob design, so during the volume detachment, if there are more than 100 missed schedules, then it does not start the job [[ref](CronJob limitations)]. In v1.1.2, the problem is resolved by reactivating the recurring job when the volume gets attached every time.
- Improve replica rebuilding speed
- In 1.1.2, speeding up replica rebuilding is by improving the performance of replica file extent retrieval by adopting the concurrent processing way instead of iterative retrieval. This can bring us several times of improvement. [ref]
- Deprecate Volume Attachement Recovery policy
- Before v1.1.2, Longhorn can manage volume attachment in different cases, but this function conflicts with the Pod Deletion Policy feature which lets kubelet handle the volume attachment lifecycle instead. To avoid this needless possible conflicting behavior, it is deprecated in v1.1.2.
- And more, please check the below list.
Installation
Longhorn supports 3 installation ways including Rancher catalog, Kubectl, and Helm. Follow the installation instructions here.
Upgrade
Follow the upgrade instructions here.
Highlights
- [ENHANCEMENT] Speed up rebuilding by getting checksum simultaneously (1948) - @joshimoo @khushboo-rancher
- [BUG] Volume Attachment Recovery Policy feature conflicts with Pod Deletion Policy (2060) - @joshimoo @smallteeths
- [BUG] Application couldn't mount the volume after expanded the volume (2422) - @cclhsu @khushboo-rancher
- [BUG] DR restored data is faulty in sles15-sp2 (2503) - @c3y1huang @khushboo-rancher
- [BUG] Failed rebuilding a large replica of 3TB (2507) - @joshimoo @khushboo-rancher
- [BUG] Recurring jobs never run again when the volume is attached after being detached for a long time (2513) - @kaxing @PhanLe1010
- [BUG] Volumes are not properly mounted/unmounted when kubelet restarts (2629) - @joshimoo @khushboo-rancher
Enhancements
- [ENHANCEMENT] Add labels for created cron jobs / backup pods for recurring jobs (2701) - @kaxing @PhanLe1010
- [FEATURE] Enable pprof endpoints in longhorn manager for troubleshooting (2715) - @kaxing @joshimoo
Bugs
- [BUG] Backup fails to complete due to timeout causing duplicate backups (2278) - @shuo-wu
- [BUG] /engineUpgrade API error handling (2289) - @PhanLe1010 @khushboo-rancher
- [BUG] Priority class not getting set for recurring jobs (2444) - @PhanLe1010 @khushboo-rancher
- [BUG] Backing image can't be delete if key more than 63 characters (2522) - @shuo-wu @khushboo-rancher
- [BUG] After network disconnection replicas cannot be reattached (2561) - @meldafrawi @shuo-wu
- [BUG] Name column in event table on the volume detail page should be bigger. (2567) - @smallteeths @khushboo-rancher
- [BUG] Detach button on volume detail page doesn't work for a volume attach to a pod. (2587) - @smallteeths @khushboo-rancher
- [BUG] full data over WS at per 30/seconds interval seems useless (2591) - @c3y1huang @khushboo-rancher
- [BUG] UI mutiple selections should be released after performing action (2595) - @smallteeths @khushboo-rancher
- [BUG] Unable to provide own tls certificate for ingress using helm chart (2615) - @kaxing @jenting
- [BUG] Failure to create a volume on a retry since v1.1.1 (2624) - @PhanLe1010 @khushboo-rancher
- [BUG] Fix missed deepcopy cases for Volume.Status.KubernetesStatus and EngineStatus.Snapshots.Snapshot (2631) - @joshimoo
- [BUG] Normalize pod order for v.status.KubernetesStatus updates (2632) - @joshimoo @khushboo-rancher
- [BUG] Fix AnnotateAWSIAMRoleArn modifying lister returned pod instances directly (2633) - @joshimoo @khushboo-rancher
- [BUG] Volume stuck in Detaching when detached during rebuild / snapshot coalescing (EXT4) (2640) - @joshimoo
- [BUG] List order caused full data transfer when no resource updated. (2646) - @c3y1huang @khushboo-rancher
- [BUG] UI multiple selections should be released after performing action on the backup page (2652) - @kaxing @smallteeths
- [BUG] List order caused full backup image data transfer when no resource updated. (2672) - @kaxing @c3y1huang
- [BUG] kubectl drain node gets stuck forever (2673) - @cclhsu @khushboo-rancher
- [BUG] Reduce Node Status updates by truncating disk StorageAvailable (2680) - @kaxing @joshimoo
- [BUG] Improve usage of Readonly listers where we only read (2681) - @joshimoo
- [BUG]NPE when listing backups of a backup volume (2683) - @kaxing @jenting
- [BUG] volume creation with an invalid FromBackup should fail (2686) - @kaxing @joshimoo
- [BUG] Liveness/Readiness Probe timeoutSeconds at Kubernetes v1.20+ (2699) - @kaxing @jenting
- [BUG] Backup / restore with AWS IAM Role fails with NoCredentialProviders (2718) - @jenting