github-actions[bot] commented on code in PR #63312:
URL: https://github.com/apache/doris/pull/63312#discussion_r3253002196
##########
be/src/exec/scan/task_executor/time_sharing/prioritized_split_runner.cpp:
##########
@@ -99,9 +110,18 @@ Result<SharedListenableFuture<Void>>
PrioritizedSplitRunner::process() {
_wait_nanos.fetch_add(start_nanos - _last_ready.load());
auto process_start_time = _ticker->read();
Review Comment:
This new split-runner boundary catches `doris::Exception`, but it does not
enable `enable_thread_catch_bad_alloc` while `process_for()` runs. If this
fallback is meant to be the generic worker boundary, allocator/memory-limit
failures inside a split will not be thrown as `MEM_ALLOC_FAILED` and therefore
will not reach `_finished_future` as an error. Please use the existing catch
macro pattern or add the same guard here.
##########
be/src/exec/scan/scanner_scheduler.cpp:
##########
@@ -75,10 +75,12 @@ Status
ScannerScheduler::submit(std::shared_ptr<ScannerContext> ctx,
TabletStorageType type = scanner_delegate->_scanner->get_storage_type();
auto sumbit_task = [&]() {
auto work_func = [scanner_ref = scan_task, ctx]() {
- auto status = [&] {
- RETURN_IF_CATCH_EXCEPTION(_scanner_scan(ctx, scanner_ref));
- return Status::OK();
- }();
+ Status status = Status::OK();
+ try {
Review Comment:
This replaces `RETURN_IF_CATCH_EXCEPTION`, but the macro also increments
`enable_thread_catch_bad_alloc` for the duration of `_scanner_scan`. Without
that guard, Doris allocator checks in `allocator.cpp` take the `no throw
exception` path for memory-tracker/sys-memory failures, so a scanner that hits
`MEM_ALLOC_FAILED` can continue instead of returning a failed scan task. Please
keep the macro (or otherwise preserve the guard and MEM_ALLOC_FAILED
conversion) around the scan work.
--
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]
