Github user paul-rogers commented on a diff in the pull request:
https://github.com/apache/drill/pull/822#discussion_r118813748
--- Diff:
exec/java-exec/src/main/java/org/apache/drill/exec/physical/impl/aggregate/HashAggTemplate.java
---
@@ -306,14 +685,29 @@ public AggOutcome doWork() {
while (true) {
// Cleanup the previous batch since we are done processing it.
+ long pre = allocator.getAllocatedMemory();
for (VectorWrapper<?> v : incoming) {
v.getValueVector().clear();
}
+ long beforeAlloc = allocator.getAllocatedMemory();
+
+ // Get the next RecordBatch from the incoming
IterOutcome out = outgoing.next(0, incoming);
+
+ // If incoming batch is bigger than our estimate - adjust the
estimate
+ long afterAlloc = allocator.getAllocatedMemory();
+ long incomingBatchSize = afterAlloc - beforeAlloc;
+ if ( /* ! handlingSpills && */ estMaxBatchSize <
incomingBatchSize ) {
+ logger.trace("Found a bigger incoming batch: {} , prior
estimate was: {}",incomingBatchSize,estMaxBatchSize);
+ estMaxBatchSize = incomingBatchSize;
+ }
+
if (EXTRA_DEBUG_1) {
logger.debug("Received IterOutcome of {}", out);
}
switch (out) {
+ case RESTART:
+ logger.warn("HASH AGG: doWork got a RESTART...");
--- End diff --
Shouldn't this throw an `IllegalStateException`? Something is terribly
wrong if we get this where it is not expected. Simply logging zillions of
messages to the log is not likely to call attention to the problem.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
