Adds support for a Google Cloud SQL sidecar #297
Conversation
|
|
|
Hey @oddeirikigland , thank you for your contribution! |
Of course, added a short explanation now @itaiad200 |
charts/lakefs/README.md
Outdated
| @@ -130,5 +150,7 @@ helm upgrade -f my-values.yaml my-lakefs lakefs/lakefs --set kv_upgrade=true | |||
| | `s3Fallback.enabled` | If set to true, an [S3Proxy](https://github.com/gaul/s3proxy) container will be started. Requests to lakeFS S3 gateway with a non-existing repository will be forwarded to this container. | | | |||
| | `s3Fallback.aws_access_key` | An AWS access key to be used by the S3Proxy for authentication | | | |||
| | `s3Fallback.aws_secret_key` | An AWS secret key to be used by the S3Proxy for authentication | | | |||
| | `gcpFallback.enabled` | If set to true, an [GCP Proxy](https://github.com/GoogleCloudPlatform/cloud-sql-proxy) container will be started. | | | |||
| | `gcpFallback.instances` | The instances to connect to. | | | |||
There was a problem hiding this comment.
instances means there could be more than a single value right? How do you expect them to be passed?
There was a problem hiding this comment.
The proxy supports several instances, that is why. One database instance is enough for this case, so I can change it to instance for simplification.
I've tested it by adding a test file to the ci folder, enabling the gcpFallback with a path to my SQL instance in GCP. Also, I've used this setup for my self-hosted lakefs deployment in GCP.
it would be great to have a charts/lakefs/ci/gcp-fallback-values.yaml, but then an open SQL gcp instance is needed. The config will then look as follows:
gcpFallback:
enabled: true
instance: <my_project>:<my_region>:<my_instance>=tcp:5432
serviceAccount:
name: <my_service_account>
closes #298