ClickHouse / ClickHouse/clickhouse-operator
Multi K8S cluster/region support
- Dominant language
- Go
- Stars
- 292
- Forks
- 42
- Avg merge
- 10h 18m
- Merged PRs (30d)
- 23
Description
### Company or project name
_No response_
### Use case
Creating a multi K8S cluster setup, where clusters may be located via sufficiently fast/close links for redundancy in case of complete datacenter failure for self hosters.
### Describe the solution you'd like
Allow to create a base cluster and attach other clusters to it with separate deployments for clickhouse-keeper and clickhouse-server. Right now using the legacy bitnami charts, I can do such setup by overriding the cluster config and specifying all instances.
EG
```xml
0
clickfuse-clickhouse-keeper-0.clickfuse-clickhouse-keeper-headless.langfuse.svc.k8s-cluster.datacenter1
1
clickfuse-clickhouse-keeper-1.clickfuse-clickhouse-keeper-headless.langfuse.svc.ik8s-cluster.datacenter1
2
clickfuse-clickhouse-keeper-2.clickfuse-clickhouse-keeper-headless.langfuse.svc.k8s-cluster.datacenter1
3
clickfuse-clickhouse-keeper-0.clickfuse-clickhouse-keeper-headless.langfuse.svc.ik8s-cluster.datacenter2
4
clickfuse-clickhouse-keeper-1.clickfuse-clickhouse-keeper-headless.langfuse.svc.k8s-cluster.datacenter2
5
clickfuse-clickhouse-keeper-2.clickfuse-clickhouse-keeper-headless.langfuse.svc.k8s-cluster.datacenter2
6
clickfuse-clickhouse-keeper-0.clickfuse-clickhouse-keeper-headless.langfuse.k8s-cluster.datacenter3
```
and a similar config for clickhouse-server
### Describe alternatives you've considered
Using the old bitnami charts but upgrading images and adapting scripts
### Additional context
Clickhouse deployment used for langfuse
Contributor guide
No contributing guide indexed for this repository
Research direction
Start by tracing how the operator configures clickhouse-keeper and clickhouse-server deployments and cluster settings; no source files or tests are named in the issue. Compare the requested base-cluster and attached-cluster model with the legacy Bitnami chart configuration shown. Done means supporting separate deployments across Kubernetes clusters or regions with the required cross-cluster Keeper and server configuration.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- clickhouse, go, kubernetes
- Domain
- databases, distributed-systems, infrastructure
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100