Hi all,

I wanted to share a new PR that addresses a bug where the "kill query"
command fails to actually terminate the running query.

**Why we need this:**
Previously, issuing a kill command wouldn't properly stop the execution,
which could lead to runaway queries tying up system resources. This is
especially problematic if a user needs to cancel a heavy, long-running
query, such as those executed in Table Mode.

**What changed:**
This PR updates the query state machine and coordinator to ensure
cancellation signals are correctly propagated. We've fixed the handling
inside the RPC service and query context so that when a kill command is
received, the execution engine properly halts the fragment instances,
updates the query execution state, and cleans up the resources.

The PR link is: https://github.com/apache/iotdb/pull/17358

Best regards,
--------------------
Yuan Tian

Reply via email to