coproc: Server side check for multiple scripts producing on exact same materialized topics

[graphcareful]

A materialized topic schema is <input_topic>.$<destination_topic>$ , so if two different coprocessors produce on the same destination_topic for which those records originated from the same input_topic, those two writes will race to the same underlying log and crash redpanda.

Long term solution would be to namespace the materialized_topic schema with the script_id.

Short term proposal is to perform a possibly expensive check before writing to a materialized log

[emaxerrno]

Hmm. I don't follow. This is a bug on coproc then. Why would semantics change

[graphcareful]

Gotcha, this is expected behavior, closing this and opening a bug report instead