## [Release 1.3.0]

There are a few improvements since last release v1.2.0.

#### Break changes:
- backup TidbCluster’s metadata to pd/tikv’s data directory. #307 @weekface (TiDB v2.0.4 required)

#### Main:
- remove tolerations of tikv/pd pod, and add a new pod when pd/tikv pod’s node is unreachable. #309 @xiaojingchen
- fix volume-manager's compatibility with K8s v1.10. #319 @weekface
- avoid tikvset's lastHeartbeatTime overwritten by zero. #328, #333 @xiaojingchen
- add tidb-lightning support. #321 @weekface (TiDB v2.0.4 required)

#### Chart:
- set volume-manager to optional, and add a `serviceExternalTrafficPolicy` parameter. #306 @weekface
- add `storageclasses` resources to `kube-scheduler`’s cluster role, #311 @weekface
- save grafana config as a configmap. #301, #329 @xiaojingchen
- add alertmanagerURL to values.yaml. #305 @xiaojingchen
- update alert.rules to latest. #330 @weekface

#### Docs:
- update README.md to reflect the latest user guide. #308 @weekface , #310 @CaitinChen
- add a CONTRIBUTING guide and Makefile commands. #313, #314 @gregwebs

#### Log:
- log the desired volume quantity. #317 @gregwebs
- create an errors package that logs and traces. #322 @gregwebs

#### Tools:
- highlight "make check" with backtick quote. #316 @gregwebs
- re-use testing script, always use ginkgo. #320 @gregwebs
- enable vet-shadow. #323 @gregwebs
- expose the e2e test through the Makefile. #326 @gregwebs
- add make clean to the ci build. #332 @gregwebs

#### Bug or typo fix:
- #312 @weekface
- #318 @xiaojingchen
- clear tikvset status when pods are all deleted. #327 @weekface

#### Other:
- fix goroutine error handling. #325 @gregwebs



## [Release 1.0.1]

#### Fixed

- Fix maximum delay of 2 minutes of pv mount by bound PV with PVC before schedulering pod