andygrove opened a new pull request, #2741: URL: https://github.com/apache/datafusion-ballista/pull/2741
# Which issue does this PR close? N/A # Rationale for this change `ballista.execution.mode` offers `blocking` and `pipelined`. "Blocking" describes the exchange from the operator's point of view and is easy to confuse with blocking I/O or blocking operators such as `SortExec`. Spark calls the same model staged execution, and Ballista already says "stage" everywhere else, so `staged` is the clearer name. # What changes are included in this PR? - `ExecutionMode::Blocking` is now `ExecutionMode::Staged`, and the `ballista.execution.mode` value `blocking` is now `staged` (the new default). The old value is not kept as an alias. - Comments and tests that refer to the mode were updated, including renamed test functions in `client/tests/pipelined.rs` and `cache_linker.rs`. - Docs were updated to match: FAQ, shuffle and architecture pages, benchmarking pages, and the DataFusion Distributed comparison. The generated configuration reference was regenerated with `./dev/update_config_docs.sh`. - Unrelated uses of "blocking" (`spawn_blocking`, `blocking_send`, "non-blocking", blocking operators, and test helpers such as `BlockingEngine`) are unchanged. # Are there any user-facing changes? The config value `blocking` is renamed to `staged`. This is not a breaking change because the blocking/pipelined split has not been in a release yet. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
