Details
-
Epic
-
Status: Resolved
-
Minor
-
Resolution: Fixed
-
Impala 2.6.0
-
Impala Multi-threaded query execution
Description
Currently, a single query fragment is run in a quasi-single threaded manner on a node: the scanners are run in multiple threads, but all other operators (joins, aggregation) are run in the main thread.
The goal is to add multi-threaded execution on a single node by running multiple fragment instances (each of which runs in a single thread).
Attachments
Attachments
Issue Links
- is a parent of
-
IMPALA-3567 Factor out join build-side
-
- Resolved
-
- is depended upon by
-
IMPALA-9807 Multi-threaded query execution - Further Enhancements
-
- Open
-
-
IMPALA-8965 Cleanup after switch to multi-threaded query execution
-
- Open
-
- is duplicated by
-
IMPALA-2849 Investigate on improving the parallelism on a single node
-
- Resolved
-
- is related to
-
IMPALA-753 Parallelize HBase scan
-
- Resolved
-
- relates to
-
IMPALA-10486 Multithreading upgrade path for large clusters
-
- Open
-
Issues in epic
|
IMPALA-4080 | Share codegen work between fragment instances |
|
Resolved | Bikramjeet Vig | |
|
IMPALA-4224 | Add backend support for join build sinks in parallel plans |
|
Resolved | Tim Armstrong | |
|
IMPALA-4400 | Aggregate runtime filters locally |
|
Resolved | Tim Armstrong | |
|
IMPALA-6267 | MT Scanners do not check runtime filters per-file before processing each split |
|
Resolved | Tim Armstrong | |
|
IMPALA-8125 | Limit number of files generated by insert |
|
Resolved | Bikramjeet Vig | |
|
IMPALA-9000 | Fix all the TODO-MT comments |
|
Resolved | Tim Armstrong | |
|
IMPALA-9099 | Allow setting mt_dop manually for queries with joins |
|
Resolved | Tim Armstrong | |
|
IMPALA-9126 | Cleanly separate build and probe state in hash join node |
|
Resolved | Tim Armstrong | |
|
IMPALA-9127 | Clean up probe-side state machine in hash join |
|
Resolved | Tim Armstrong | |
|
IMPALA-9156 | Share broadcast join builds between fragments |
|
Resolved | Tim Armstrong | |
|
IMPALA-9176 | Make access to null-aware partition from PartitionedHashJoinNode read-only |
|
Resolved | Tim Armstrong | |
|
IMPALA-9197 | Hash table lookup should be read-only |
|
Resolved | Tim Armstrong | |
|
IMPALA-9302 | Multithreaded scanners don't check for filter effectiveness |
|
Resolved | Tim Armstrong | |
|
IMPALA-9317 | Improve number of instances estimate for scans in planner |
|
Resolved | Tim Armstrong | |
|
IMPALA-9318 | Guard rail for mt_dop value |
|
Resolved | Joe McDonnell | |
|
IMPALA-9422 | Improve join builder profiles |
|
Resolved | Bikramjeet Vig | |
|
IMPALA-9586 | Update query option docs to account for interactions with mt_dop |
|
Resolved | Tim Armstrong | |
|
IMPALA-9782 | KuduPartitionExpr is not thread-safe |
|
Resolved | Tim Armstrong | |
|
IMPALA-10343 | control_service_queue_mem_limit default is too low for large clusters |
|
Resolved | Tim Armstrong | |
|
IMPALA-10351 | Enable mt_dop for DML |
|
Resolved | Tim Armstrong |