Must specify a schema config when using S3 block storage

[RyanW8]

With the below configuration I am unable to generate the kubernetes manifests. I get this error evaluating jsonnet: RUNTIME ERROR: must specify a schema config. My understanding was that when using s3 block storage it wasn't required to specify a schema?

local cortex = import "cortex/cortex.libsonnet";

  _config+:: {
    namespace: "cortex",
    blocks_storage: {
      backend: 's3',
      tsdb: {
        dir: '/data/cortex-tsdb-ingester',
        ship_interval: '1m',
        block_ranges_period: [ '2h' ],
        retention_period: '3h'
      },
      bucket_store: {
        sync_dir: '/data/cortex-tsdb-querier'
      },
      s3: {
        bucket_name: 'cortex-tsdb',
        endpoint: 'minio-minio-sre-s3-minio.minio.svc.cluster.local:9000',
        secret_access_key: 'REDACTED',
        access_key_id: 'n4monitoring'
      }
    },
    ingester: {
      statefulset_replicas: 3,
      max_transfer_retries: 0,
      lifecycler: {
        join_after: '0s',
        final_sleep: '0s',
        num_tokens: 512,
        ring: {
          replication_factor: 1,
          kvstore: {
            store: 'etcd',
            prefix: 'collectors/',
            etcd: {
              endpoints: [
                'cortex-etcd.cortex.svc.cluster.local:2379'
              ]
            }
          }
        }
      }
    },
    storage: {
      engine: 'blocks'
    },
    store_gateway: {
      sharding_enabled: true,
      sharding_ring: {
        kvstore: {
          store: 'etcd',
          etcd: {
            endpoints: [
              'cortex-etcd.cortex.svc.cluster.local:2379'
            ]
          }
        }
      }
    }
  },
}

[pstibrany]

Hi, _config isn't just plan translation of Cortex configuration YAML file. It has its own structure, see https://github.com/grafana/cortex-jsonnet/blob/master/cortex/config.libsonnet. To enable blocks storage, you need to use storage_engine: 'blocks' inside _config object (done automatically if you also import "cortex/tsdb.libsonnet" and add it to the config:

local cortex = import 'cortex/cortex.libsonnet';
local tsdb = import 'cortex/tsdb.libsonnet';

cortex + tsdb {
  _config+:: {
    cluster: 'cluster-name',
    namespace: 'cortex',

    blocks_storage_bucket_name: 'bucket_name',
    cortex_store_gateway_data_disk_size: '10Gi',
  },
}

Another example (using chunks storage) is at https://github.com/grafana/cortex-jsonnet/blob/master/cortex/cortex-manifests.jsonnet.example