Component version updates
- Linstor: `v1.19.1`
- DRBD: `v9.1.8`
- drbd-utils: `v9.21.4`
- drbd-reacotr: `v0.8.0`
- linstor-csi: `v0.20.0`
- linstor-scheduler-extender: `0.2.1`
- linstor-server: `v1.19.1`
- linstor-client: `v1.14.0`
- linstor-api: `v1.14.0`
- piraeus-ha-controller: `v1.1.0`
- piraeus-operator: `v1.9.1`
- Prometheus: `v2.37.0`
- CSI components
- provisioner:` v3.2.1`
- attacher: `v3.5.0`
- resizer: `v1.5.0`
- registrar: `v2.5.1`
- snapshotter: `v6.0.1`
- livenessprobe: `v2.7.0`
Important update notes
- All linstor components will be moved from master to system nodes.
- Ingress controllers will restart.
- Prometheus Pods will be restarted.
- Some system Pods will be restarted: `kube-dns`, `chrony`, Pods of cni-* modules and cloud-provider-* modules.
- Webhook handler will restart. During the handler restart, Deckhouse controller could generate a few error messages when it will not be able to access the webhook. It should be resolved in the next 15 seconds.
Major changes
- Prometheus has been updated to LTS version 2.37.0. Reduced memory consumption.
- Linstor update with a lot of changes.
- Added the ability to specify a master passphrase.
- Improved backup functionality (including on S3).
- CSI components have been updated.
- Deckhouse checks the version of the Ingress controller when updating the release. If the version of the Ingress controller used in the cluster does not meet the requirements, the Deckhouse release update will not be executed (until the problem is corrected).
- Ability to disable metrics collection from specific Namespaces and Ingress resources.
- Control of disruptive updates. The deckhouse.update.disruptionApprovalMode configuration parameter allows enabling confirmation mode for Deckhouse updates that could explicitly (or potentially) cause the Kubernetes cluster payload (user applications) to become unavailable.
See [CHANGELOG v1.35](https://github.com/deckhouse/deckhouse/blob/main/CHANGELOG/CHANGELOG-v1.35.md) for more details.