Describe what's wrong
Attempting to join a remote table (S3Cluster) to a local table and getting an error

How to reproduce

• Have a multinode clickhouse cluster
• Create a table that's distributed
• S3 bucket with a parquet file
• Try to query the distributed table with a JOIN using s3Cluster wrapper for remote data source

Which ClickHouse server version to use

• 23.6.1.1524

Queries to run that lead to unexpected result
Notes:

• Parquet file
• s3 bucket with a parquet file. any data model should cause the error
• clickhouse cluster with > 1 node
• Both with and without CTE breaks

** If you flip the tables and join the local table into the remote table it works. It only fails if the local table comes before the remote table

WITH some_remote_table AS (SELECT * FROM s3Cluster('<clickhouse cluster name>', '<s3 object url>', '<s3 object access key>', '<s3 object person key>', 'Parquet')) SELECT some_local_table.event FROM some_local_table JOIN some_remote_table ON some_local_table.event = some_remote_table.id

Expected behavior

Query should return

Error message and/or stacktrace

Distributed task iterator is not initialized