We've just released a release candidate for the next Kubernetes Operator 2.11.0.
It contains a number of improvements that warrant a longer release
phase for testing.
The Operator itself gained support for limiting the number of
concurrent evacuations using the new
LinstorCluster.spec.maxConcurrentEvacuations key. It also now allows
passing arguments to lvcreate, zpool create, etc... on storage pool
creation.
On the new component feature front, we have added support for
consistency groups in volumes. This ensures that in the event of node
failures, a workload using multiple volumes at once sees a consistent
state of the data on the new host after failover. [1]
We also implemented optional Quality of Service settings for volumes.
These require an NRI plugin that enforces the necessary limits on Pod
startup. [2]
We also have the usual round of improvements and bug fixes by updating
our components. LINSTOR itself should be faster to start up on first
deployment. LINSTOR CSI should be more robust when validating RWX
modes, along with the usual dependency updates.
We have also updated the DRBD release to 9.3.3, making this the first
release using the 9.3 branch. This means we will no longer support
RHEL7 and clones for our Operator going forward.
To upgrade, either point your kustomization.yaml at the new manifest:
https://charts.linstor.io/static/v2.11.0-rc.1.yaml
Or, if using helm, upgrade the linstor-operator chart:
helm repo update
helm upgrade linstor-operator linstor/linstor-operator --wait
--version 2.11.0-rc.1
To get specific instructions to apply the update, check our users guide[3].
Source code is, as always, available upstream[4].
Please help by testing this release!
Best regards,
Moritz
[1]: Use Consistency Groups - Piraeus Datastore
[2]: Configure Volume IO Limits (QoS) - Piraeus Datastore
[3]: LINSTOR 1.0 en - LINBIT
[4]: GitHub - piraeusdatastore/piraeus-operator: The Piraeus Operator manages LINSTOR clusters in Kubernetes. · GitHub
---
### Added
- Allow configuring the separator between generated Satellite resource
names and the Satellite name.
- Add `LinstorCluster.spec.maxConcurrentEvacuations` to limit how many
Satellites are evacuated at the same time, for both direct and
ClusterAPI-driven node removals.
- Enable consistency groups in LINSTOR CSI: PersistentVolumeClaims
sharing a `linstor.csi.linbit.com/consistency-group` label are placed
as separate volumes of one LINSTOR resource, ensuring a fail-over node
always sees a consistent view across the whole set of volumes.
### Changed
- Updated images:
* LINSTOR 1.34.1
* LINSTOR CSI 1.12.0
* LINSTOR Affinity Controller 1.5.0
* DRBD 9.3.3
* DRBD Reactor 1.12.0