github-actions[bot] commented on code in PR #69566:
URL: https://github.com/apache/doris/pull/69566#discussion_r4366239544
##########
fe/fe-core/src/main/java/org/apache/doris/qe/MysqlConnectProcessor.java:
##########
@@ -281,6 +285,21 @@ protected void handleExecute(PrepareCommand
prepareCommand, long stmtId, Prepare
}
}
+ /**
+ * COM_STMT_EXECUTE has no per-statement close like the one COM_QUERY runs
in its finally. An execution that
+ * fails after its scans were planned, such as one a scan-level SQL block
rule rejects or a prepared
+ * INSERT ... SELECT whose transaction fails to begin, never reaches
unregisterQuery: the query-finish callbacks
+ * its scans registered (releasing a connector read transaction, closing
the connector scope) would stay in the
+ * registry for good, its connector scope would stay open until the
statement's next EXECUTE, and scan nodes left
+ * for the statement end would never stop. The context ExecuteCommand ran
this execution with is the prepared
+ * statement's current one (see
PreparedStatementContext#nextStatementContext). Both calls are no-ops for an
+ * execution that finished normally, which already unregistered its query
and closed its scope.
+ */
+ private void finishExecution(PreparedStatementContext prepCtx) {
+ QeProcessorImpl.INSTANCE.unregisterQuery(ctx.queryId());
Review Comment:
[P2] Wait for split producers before closing the execution scope
An Iceberg streaming scan can hit the first-split timeout while its producer
is already running but has not reached the statement-scope lookup; for example,
scheduler backlog can start it just before the 30-second deadline.
`unregisterQuery()` runs the scan's stop callback and then
`QueryFinishCleanup`, which calls `statementScope.closeAll()`, but
`SplitProducer.close()` only requests cancellation and does not wait for that
running producer. The producer can therefore resume after the scope was cleared
and insert a new `ScopedBorrow`/tracked table through `computeIfAbsent` on the
already-closed scope. Its split source closes only task iterators, while
subsequent `closeAll()` calls (including the next EXECUTE's reset) return
immediately because `closed` is already true, permanently retaining the
table/FileIO/catalog lease. Previously, if the producer finished before the
next EXECUTE, the still-open outgoing scope released that holder at reset.
Quiesce producers before closing their
scope, or defer scope teardown until they complete, and cover timeout versus
in-flight scope lookup.
--
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]
