Longhorn v0.6.0 Release

Longhorn v0.6.0 is an alpha-quality release.

Longhorn v0.6.0 is a major release with multiple high-impact features. We had made a major architecture change in v0.6.0 and won't be able to support live upgrade from v0.5.0. The offline upgrade from v0.5.0 is supported.

Highlights:

1. Speed up volume attaching/detaching process by 50% or more. Longhorn will create an Instance Manager on each node to manage Longhorn Engine processes. No more one Pod per Engine/Replica. Now the setting GuaranteedEngineCPU will apply to the Instance Manager Pod instead.
2. Storage Tag support. Now the user can specify which disk/node they want the volume to scheduled in by using the tag e.g. hdd, ssd.
3. Hard affinity support. Now the user can stop Longhorn from using the same node for rebuilding the replica even there is no other node available.
4. Support taint toleration. Now the user can dedicate certain nodes for Longhorn components and avoid putting other workloads on it using Kubernetes taint.
5. Support labels for backups. Now the user can specify labels for certain backups. It works for scheduled backups as well. And Longhorn will automatically tag backup if the volume was used for a running workload.
6. Support marking nodes with certain labels as disabled by default. Now the user can configure Longhorn to create the default data path only for the certain nodes with the labels, so Longhorn will not automatically use all the newly added nodes.
7. Support set default data path for storing data. Now the default path for storing data when adding a new node is customizable.
8. Customizable setting when installing Longhorn. Now Longhorn settings can be configured as a part of the installation.
9. Automatically attach DR and newly restored volumes. Now it's no longer necessary to manually attach the volume to trigger the restoration process.
10. Support Pod Security Policy.
11. UI now shows the progress for backup, restore and snapshot deleting.

Note:

1. The base image feature was removed in v0.6.0. If you're using the feature, do not upgrade to v0.6.0.

Longhorn v0.6.0-rc2 release

This is the second Release Candidate(RC2) for Longhorn v0.6.0 release. Do not upgrade to Release Candidate releases.

Longhorn v0.6.0 will be an alpha-quality release.

Longhorn v0.6.0 is a major release with multiple high-impact features. We had made a major architecture change in v0.6.0 and won't be able to support live upgrade from v0.5.0. The offline upgrade from v0.5.0 is supported.

Highlights:

1. Speed up volume attaching/detaching process by 50% or more. Longhorn will create an Instance Manager on each node to manage Longhorn Engine processes. No more one Pod per Engine/Replica. Now the setting GuaranteedEngineCPU will apply to the Instance Manager Pod instead.
2. Storage Tag support. Now the user can specify which disk/node they want the volume to scheduled in by using the tag e.g. hdd, ssd.
3. Hard affinity support. Now the user can stop Longhorn from using the same node for rebuilding the replica even there is no other node available.
4. Support taint toleration. Now the user can dedicate certain nodes for Longhorn components and avoid putting other workloads on it using Kubernetes taint.
5. Support labels for backups. Now the user can specify labels for certain backups. It works for scheduled backups as well. And Longhorn will automatically tag backup if the volume was used for a running workload.
6. Support marking nodes with certain labels as disabled by default. Now the user can configure Longhorn to create the default data path only for the certain nodes with the labels, so Longhorn will not automatically use all the newly added nodes.
7. Support set default data path for storing data. Now the default path for storing data when adding a new node is customizable.
8. Customizable setting when installing Longhorn. Now Longhorn settings can be configured as a part of the installation.
9. Automatically attach DR and newly restored volumes. Now it's no longer necessary to manually attach the volume to trigger the restoration process.
10. Support Pod Security Policy.
11. UI now shows the progress for backup, restore and snapshot deleting.

Note:

1. The base image feature was removed in v0.6.0. If you're using the feature, do not upgrade to v0.6.0.

Longhorn v0.6.0-rc1 Release

This is the first Release Candidate(RC1) for Longhorn v0.6.0 release. Do not upgrade to Release Candidate release.

Longhorn v0.6.0 would be an alpha-quality release.

Longhorn v0.6.0 is a major release with multiple high-impact features. We had made a major architecture change in v0.6.0 and won't be able to support live upgrade from v0.5.0. The offline upgrade from v0.5.0 is supported.

Highlights:

1. Speed up volume attaching/detaching process by 50% or more. Longhorn will create an Instance Manager on each node to manage Longhorn Engine processes. No more one Pod per Engine/Replica. Now the setting GuaranteedEngineCPU will apply to the Instance Manager Pod instead.
2. Storage Tag support. Now the user can specify which disk/node they want the volume to scheduled in by using the tag e.g. hdd, ssd.
3. Hard affinity support. Now the user can stop Longhorn from using the same node for rebuilding the replica even there is no other node available.
4. Support taint toleration. Now the user can dedicate certain nodes for Longhorn components and avoid putting other workloads on it using Kubernetes taint.
5. Support labels for backups. Now the user can specify labels for certain backups. It works for scheduled backups as well. And Longhorn will automatically tag backup if the volume was used for a running workload.
6. Support marking nodes with certain labels as disabled by default. Now the user can configure Longhorn to create the default data path only for the certain nodes with the labels, so Longhorn will not automatically use all the newly added nodes.
7. Support set default data path for storing data. Now the default path for storing data when adding a new node is customizable.
8. Customizable setting when installing Longhorn. Now Longhorn settings can be configured as a part of the installation.
9. Automatically attach DR and newly restored volumes. Now it's no longer necessary to manually attach the volume to trigger the restoration process.
10. Support Pod Security Policy.
11. UI now shows the progress for backup, restore and snapshot deleting.

Note:

1. The base image feature was removed in v0.6.0. If you're using the feature, do not upgrade to v0.6.0.

Longhorn v0.5.0 release

Highlights:

1. Users now can use Disaster Recovery Volume support (#495 ) to recover the volume in another Kubernetes cluster with defined RTO and RPO. See here for details
2. Users now can see Kubernetes workload information and create PV/PVC in Longhorn UI (#461 ) .See here for details
3. Users now can set backup scheduling in the storage class (#362)
4. We now add the Helm chart in the Longhorn repo, in addition to Rancher Apps. (#445 )

See all the issues resolved in v0.5.0 at:
https://github.com/rancher/longhorn/milestone/3?closed=1

The volume engines would need to upgrade to v0.5.0 as well. Please follow the instruction to upgrade engines.

For Rancher v2.2 users with Catalog Apps, please make sure install Longhorn in longhorn-system namespace (instead of longhorn namespace). There is a recent bug in Rancher v2.2 affects the fresh installation of Longhorn App. The fix will be in Rancher v2.2.4 release.

Longhorn v0.5.0-rc3 release

1. Users now can use Disaster Recovery Volume support (#495 ) to recover the volume in another Kubernetes cluster with defined RTO and RPO.
2. Users now can see Kubernetes workload information and create PV/PVC in Longhorn UI (#461 ).
3. Users now can set backup scheduling in the storage class (#362)
4. We now add the Helm chart in the Longhorn repo, in addition to Rancher Apps. (#445 )

Longhorn v0.5.0-rc2 release

1. Users now can use Disaster Recovery Volume support (#495 ) to recover the volume in another Kubernetes cluster with defined RTO and RPO.
2. Users now can see Kubernetes workload information and create PV/PVC in Longhorn UI (#461 ).
3. Users now can set backup scheduling in the storage class (#362)
4. We now add the Helm chart in the Longhorn repo, in addition to Rancher Apps. (#445 )

Longhorn v0.5.0-rc1 release

1. Users now can use Disaster Recovery Volume support (#495 ) to recover the volume in another Kubernetes cluster with defined RTO and RPO.
2. Users now can see Kubernetes workload information and create PV/PVC in Longhorn UI (#461 ).
3. Users now can set backup scheduling in the storage class (#362 ).

Longhorn v0.4.1 release

Longhorn v0.4.1 has added support for RancherOS and K3S.

Highlights:

1. Support for RancherOS and K3S. See https://github.com/rancher/longhorn/blob/v0.4.1/docs/rancheros.md
2. Support for restoring a backup into an image file. See https://github.com/rancher/longhorn/blob/v0.4.1/docs/restore-to-file.md
3. Improve logging mechanism across the board.

See all the issues resolved in v0.4.1 at https://github.com/rancher/longhorn/milestone/10?closed=1

The volume engines would need to upgrade to v0.4.1 as well. Please follow the instruction to upgrade the engines.

Longhorn v0.4.1-rc1 release

Merge pull request #439 from yasker/work

Sync with manager

Longhorn v0.4.0 release

Longhorn v0.4.0 is a release focused on stability.

Notable updates:

1. Greatly increases Longhorn's ability to recover from Kubernetes node failure. For expected behavior, see https://github.com/rancher/longhorn/blob/v0.4/docs/node-failure.md
2. Fixes an issue ( #380 ) that may cause block device failure and remounted as read-only by the kernel when the network became unstable.
3. Added a support bundle ( #84 ) in the UI, which enable the user to download a zip file contains the related configuration/logs, to help developers supporting Longhorn.
4. Upgradable engines now would be shown in UI.
5. Node redundancy would be shown in UI. If there are multiple replicas share the same node, an icon would be shown to remind user there is a risk.
6. Fix a UI issue ( #397) that results in an unusable Backup page behind the rancher proxy.

The volume engines would need to upgrade to v0.4.0 as well. Please follow the instruction to upgrade engine.