Longhorn v0.8.0 release

Highlights:

1. Volume expansion (resize)[experimental]. This is one of the most requested features, and we finally able to implement it in the v0.8.0 release. Longhorn v0.8.0 supports offline expansion only. See here for details.
2. Automatic recovery (salvage) from faulted volume due to the engine in the isolated network. This feature will allow the faulted volumes to be automatically salvaged and immediately available for usage once the network recovered.
3. Performance improvement on attaching/detaching We have managed to dramatically improve the attach/detach performance if the users are using large numbers of volume per node. If you're using more than 30 volumes per node, you might see 100 times+ improvement vs Longhorn v0.7.0. See here for details.
4. Support k8s zone and region tags for replica placement. This will allow us to help GKE/EKS users to achieve volume high availability even in cross AZ situation.
5. Progress report for rebuild.
6. Upgrade: Live upgrade v2: reuse the instance manager for upgrade. Now Longhorn can reuse the instance manager for live-upgrade engines. The offline upgrade will still be needed from v0.7.0 to v0.8.0 to place the mechanism in place.
7. Move test cases to python 3. Python 2 is dead. Long live Python 3!

Only upgrade from v0.7.0 to v0.8.0 is supported. Please follow the offline upgrade process for the engine upgrade

The minimal Kubernetes version supported is v1.14.0.

The minimal Kubernetes version supporting volume expansion for storage class feature is v1.16.0.

Longhorn v0.8.0-rc3 release

Upgrade from/to a release candidate is not supported.

The minimal Kubernetes version supported is v1.14.0.

Highlights:

1. Volume expansion (resize)[experimental]. This is one of the most requested features, and we finally able to implement it in the v0.8.0 release. Notice that Kubernetes only start to support volume expansion for storage class in v1.16.0. Longhorn v0.8.0 supports offline expansion only. See here for details.
2. Automatic recovery (salvage) from faulted volume due to the engine in the isolated network. This feature will allow the faulted volumes to be automatically salvaged and immediately available for usage once the network recovered.
3. Performance improvement on attaching/detaching We have managed to dramatically improve the attach/detach performance if the users are using large numbers of volume per node. If you're using more than 30 volumes per node, you might see 100 times+ improvement vs Longhorn v0.7.0. See here for details.
4. Support k8s zone and region tags for replica placement. This will allow us to help GKE/EKS users to achieve volume high availability even in cross AZ situation.
5. Progress report for rebuild.
6. Upgrade: Live upgrade v2: reuse the instance manager for upgrade. Now Longhorn can reuse the instance manager for live-upgrade engines. Offline upgrade will still be needed from v0.7.0 to v0.8.0 to place the mechanism in place.
7. Move test cases to python 3. The Python 2 is dead. Long live the Python 3!

Longhorn v0.8.0-rc2 Release

Upgrade from/to a release candidate is not supported.

The minimal Kubernetes version supported is v1.14.0.

Highlights:

1. Volume expansion (resize)[experimental]. This is a long-requested feature, and we finally able to implement it in v0.8.0. Notice that Kubernetes only start to support volume expansion for storage class in v1.16.0. See here for details.
2. Automatic recovery (salvage) from faulted volume due to the engine in the isolated network. This feature will allow the faulted volumes to be automatically salvaged and immediately available for usage once the network recovered.
3. Performance improvement on attaching/detaching We have managed to dramatically improve the attach/detach performance if the users are using large numbers of volume per node. If you're using more than 30 volumes per node, you might see 100 times+ improvement vs Longhorn v0.7.0. See here for details.
4. Support k8s zone and region tags for replica placement. This will allow us to help GKE/EKS users to achieve volume high availability even in cross AZ situation.
5. Progress report for rebuild.
6. Upgrade: Live upgrade v2: reuse the instance manager for upgrade. Now Longhorn can reuse the instance manager for live-upgrade engines. Offline upgrade will still be needed from v0.7.0 to v0.8.0 to place the mechanism in place.
7. Move test cases to python 3. The Python 2 is dead. Long live the Python 3!

Longhorn v0.8.0-rc1 release

Upgrade from/to a release candidate is not supported.

The minimal Kubernetes version supported is v1.14.0.

Highlights:

1. Volume expansion (resize)[experimental]. This is a long-requested feature, and we finally able to implement it in v0.8.0. Notice that Kubernetes only start to support volume expansion for storage class in v1.16.0. See here for details.
2. Automatic recovery (salvage) from faulted volume due to the engine in the isolated network. This feature will allow the faulted volumes to be automatically salvaged and immediately available for usage once the network recovered.
3. Support k8s zone and region tags for replica placement. This will allow us to help GKE/EKS users to achieve volume high availability even in cross AZ situation.
4. Progress report for rebuild.
5. Upgrade: Live upgrade v2: reuse the instance manager for upgrade. Now Longhorn can reuse the instance manager for live-upgrade engines. Offline upgrade will still be needed from v0.7.0 to v0.8.0 to place the mechanism in place.
6. Move test cases to python 3. The Python 2 is dead. Long live the Python 3!

Longhorn v0.7.0 release

Longhorn v0.7.0 is the first beta release of Longhorn. And it's the first release after Longhorn joined CNCF.

The minimal Kubernetes version supported is v1.14.0.

Longhorn CRD API Version has been upgraded from longhorn.rancher.io/v1alpha1 to longhorn.io/v1beta1.

Highlights:

1. Automatically reattach volume if the detaching is caused by the node reboots, Docker restarts, Kubernetes upgrades, etc. See here for details.
2. Container Storage Interface (CSI) v1.1.0, with raw block volume support.
3. K3S v0.10.0+. For K3S < v0.10.0, please refer to the CSI configuration doc.
4. NFSv4.1 support.

Upgrade:

1. Only upgrading from Longhorn v0.6.2 is supported.
   1. For other Longhorn version users, please upgrade to v0.6.2 first before upgrading to v0.7.0.
   2. The upgrade instruction is available here.

Longhorn v0.7.0-rc2 release

Longhorn v0.7.0 will be the first beta release of Longhorn. And it's the first release after Longhorn joined CNCF.

Upgrade from/to a release candidate is not supported.

The minimal Kubernetes version supported is v1.14.0.

Longhorn CRD API Version has been upgraded from longhorn.rancher.io/v1alpha1 to longhorn.io/v1beta1.

Highlights:

1. Automatically reattach volume if the detaching is caused by the node reboots, Docker restarts, Kubernetes upgrades, etc. See here for details.
2. Container Storage Interface (CSI) v1.1.0, with raw block volume support.
3. K3S v0.10.0+. For K3S < v0.10.0, please refer to the CSI configuration doc.
4. NFSv4.1 support.

Upgrade:

1. Only upgrading from Longhorn v0.6.2 is supported.
   1. For other Longhorn version users, please upgrade to v0.6.2 first before upgrading to v0.7.0.
   2. The upgrade instruction is available here.

Longhorn v0.7.0-rc1 release

Longhorn v0.7.0 will be the first beta release of Longhorn. And it's the first release after Longhorn joined CNCF.

Upgrade from/to v0.7.0-rc1 is not supported.

The minimal Kubernetes version supported is v1.14.0.

Longhorn CRD API Version has been upgraded from longhorn.rancher.io/v1alpha1 to longhorn.io/v1beta1.

Highlights:

1. Automatically reattach volume if the detaching is caused by the node reboots, Docker restarts, Kubernetes upgrades, etc.
2. Container Storage Interface (CSI) v1.1.0, with raw block volume support.
3. K3S v0.10.0+. For K3S < v0.10.0, please refer to the CSI configuration doc.
4. NFSv4.1 support.

Upgrade:

1. Only upgrading from Longhorn v0.6.2 is supported.
   1. For other Longhorn version users, please upgrade to v0.6.2 first before upgrading to v0.7.0.

Longhorn v0.6.2 release

Longhorn v0.6.2 is an alpha-quality release.

This release is for addressing the high CPU utilization issue, as well as some other issues we found since v0.6.1 release.

See the list of issues resolved here.

Please follow the offline upgrade steps to upgrade engine image to v0.6.2 to fix the high CPU utilization issue.

Longhorn v0.6.2-rc1 release

This is the first Release Candidate(RC1) for Longhorn v0.6.2 release. Do not upgrade to Release Candidate releases.

Longhorn v0.6.2 is an alpha-quality release.

This release is for addressing the high CPU utilization issue, as well as some other issues we found since v0.6.1 release.

See the list of issues resolved here.

Please follow the offline upgrade steps to upgrade engine image to v0.6.2.

Longhorn v0.6.1 release

Longhorn v0.6.1 is an alpha-quality release.

This release is for addressing the issue #754 , which results in Longhorn v0.6.0 manager failure in certain environments.

Noted: Longhorn v0.6.1 is released with Longhorn Engine v0.6.0, so the users already upgraded to v0.6.0 don't need to upgrade the engine image.