Deckhouse

Fixes


- **[node-manager]** Remove bashible-apiserver deployment to avoid race condition [2191](https://github.com/deckhouse/deckhouse/pull/2191)
- **[node-manager]** Remove race condition while updating `bashible-apiserver`. [2185](https://github.com/deckhouse/deckhouse/pull/2185)
`bashible-apiserver` will restart.


See [CHANGELOG v1.34](https://github.com/deckhouse/deckhouse/blob/main/CHANGELOG/CHANGELOG-v1.34.md) for more details.

Know before update


- After upgrading Deckhouse, all nodes with the installed `docker.io` package will request `disruptive update`. You will receive `NodeRequiresDisruptionApprovalForUpdate` if you have manual `approvalMode` in the NodeGroup.

Features


- **[candi]** Forbid use docker.io package. [2175](https://github.com/deckhouse/deckhouse/pull/2175)

Fixes


- **[candi]** Fix bash for start kubelet step. [2174](https://github.com/deckhouse/deckhouse/pull/2174)


See [CHANGELOG v1.34](https://github.com/deckhouse/deckhouse/blob/main/CHANGELOG/CHANGELOG-v1.34.md) for more details.

Know before update


- After upgrading Deckhouse all nodes with Docker 18.* will request `disruptive update`. You will receive `NodeRequiresDisruptionApprovalForUpdate` if you have manual `approvalMode` in NodeGroup.

Features


- **[candi]** Forbid using Docker 18.* [2134](https://github.com/deckhouse/deckhouse/pull/2134)
After upgrading Deckhouse all nodes with Docker 18.* will request `disruptive update`. You will receive `NodeRequiresDisruptionApprovalForUpdate` if you have manual `approvalMode` in NodeGroup.
- **[control-plane-manager]** Added feature gate `EndpointSliceTerminatingCondition` for Kubernetes 1.20. [2112](https://github.com/deckhouse/deckhouse/pull/2112)
all control-plane components should be restarted.

Fixes


- **[candi]** Use Debian buster containerd package by default. [2135](https://github.com/deckhouse/deckhouse/pull/2135)
- **[candi]** Start kubelet manually if it is not running. [2132](https://github.com/deckhouse/deckhouse/pull/2132)
- **[cni-cilium]** Bandwidth controller metrics are not erroring out now. Also added logging to three controllers so that we can diagnose possible issues better. [2155](https://github.com/deckhouse/deckhouse/pull/2155)
- **[monitoring-kubernetes]** Added alert `NodeSUnreclaimBytesUsageHigh`. [2154](https://github.com/deckhouse/deckhouse/pull/2154)

Chore


- **[docs]** Clarified docs for `rootDiskSize` in OpenStackInstanceClass [2041](https://github.com/deckhouse/deckhouse/pull/2041)


See [CHANGELOG v1.34](https://github.com/deckhouse/deckhouse/blob/main/CHANGELOG/CHANGELOG-v1.34.md) for more details.

Fixes


- **[node-local-dns]** Correct check on coredns startup. [2120](https://github.com/deckhouse/deckhouse/pull/2120)
The `node-local-dns` will restart.


See [CHANGELOG v1.34](https://github.com/deckhouse/deckhouse/blob/main/CHANGELOG/CHANGELOG-v1.34.md) for more details.

Fixes


- **[node-local-dns]** Revert service account to prevent Pod from getting stuck in a Terminating state. [2111](https://github.com/deckhouse/deckhouse/pull/2111)


See [CHANGELOG v1.34](https://github.com/deckhouse/deckhouse/blob/main/CHANGELOG/CHANGELOG-v1.34.md) for more details.

Fixes


- **[ingress-nginx]** Update the `D8IngressNginxControllerVersionDeprecated` alert description. [2088](https://github.com/deckhouse/deckhouse/pull/2088)
- **[istio]** Fixed `D8IstioDataPlanePatchVersionMismatch` alert description. [2048](https://github.com/deckhouse/deckhouse/pull/2048)
- **[monitoring-kubernetes]** Ignore containers rootfs mount point for node-exporter in GKE. [2100](https://github.com/deckhouse/deckhouse/pull/2100)
- **[prometheus]** Rollback prometheus-module ServiceMonitor label selector to `app=prometheus`. [2107](https://github.com/deckhouse/deckhouse/pull/2107)


See [CHANGELOG v1.34](https://github.com/deckhouse/deckhouse/blob/main/CHANGELOG/CHANGELOG-v1.34.md) for more details.