Impala 2.8 Change Log

New Feature

• [IMPALA-1286] - Pull out common conjuncts from disjunctions
• [IMPALA-1654] - Impala needs to support all operators in drop partitions (<, >, <>, !=, <=, >=) like hive does
• [IMPALA-1788] - Consider replacing constant exprs with literal equivalents
• [IMPALA-2521] - Introduce CLUSTERED plan hint for insert statements
• [IMPALA-2523] - Make HdfsTableSink aware of clustered input
• [IMPALA-2890] - Complete 'ALTER' DDL command support for Kudu
• [IMPALA-3725] - Support Kudu UPSERT in Impala
• [IMPALA-3973] - add position and occurrence to instr()
• [IMPALA-4000] - Restricted Sentry authorization for Kudu Tables
• [IMPALA-4531] - In-predicate not pushed to Kudu scan node
• [IMPALA-4619] - Error when specifying NULL default values in Kudu tables

Epic

• [IMPALA-2755] - Clean up memory management in backend

Improvement

• [IMPALA-641] - Allow queries to be cancelled while waiting for a resource reservation (RM).
• [IMPALA-1169] - Impala debug webpage should identify query in Admission Control queue
• [IMPALA-1570] - DROP / COMPUTE incremental stats with dynamic partition specs
• [IMPALA-2057] - Improve error message for incorrect avro decimal column declaration
• [IMPALA-2379] - Constant argument for UDAF is not accessible in Init function.
• [IMPALA-2789] - Investigate packed mem layout.
• [IMPALA-2905] - Use FragmentMgr to manage coordinator fragments
• [IMPALA-2916] - Add warning to query profile if debug build
• [IMPALA-3211] - Toolchain binary publishing improvements
• [IMPALA-3481] - Kudu Connector: use new Scan Token API
• [IMPALA-3491] - Reduce number of serial tests with unique database fixture
• [IMPALA-3704] - Fix error message when violating key constraint in Kudu insert
• [IMPALA-3718] - Improve functional testing on Impala Kudu
• [IMPALA-3788] - Add Kudu support for read-your-writes scan consistency
• [IMPALA-3823] - Add timer to measure Parquet footer reads
• [IMPALA-3884] - Enable codegen for TIMESTAMP in hash table.
• [IMPALA-3949] - CatalogServiceCatalog.extractFunctions should log the underlying exception when failing to load a function
• [IMPALA-3983] - Clean up UDF jars after reloading functions
• [IMPALA-3988] - Only use first 96 bits of query ID
• [IMPALA-4011] - "Missed topic update deadline consider increasing statestore_update_frequency_ms" log is misleading
• [IMPALA-4048] - Impala daemon web UI counts EXPIRED sessions in ACTIVE sessions
• [IMPALA-4054] - Remove serial test workarounds for IMPALA-2479.
• [IMPALA-4099] - Improve error message while loading UDFs with a missing jar
• [IMPALA-4134] - Use Kudu AUTO_FLUSH_BACKGROUND mode
• [IMPALA-4160] - Remove Llama code from Impala
• [IMPALA-4187] - Switch RPC latency metrics to histograms
• [IMPALA-4258] - Clean up duplicated test macros
• [IMPALA-4259] - Allow building Impala without thirdparty/ test dependencies
• [IMPALA-4269] - Codegen for Merging-Exchange node
• [IMPALA-4278] - Don't abort Catalog startup quickly if HMS is not present
• [IMPALA-4289] - Mark slots produced by NDV() and NDV_NO_FINALIZE() as non-nullable.
• [IMPALA-4291] - Lazy creation of IRFunction::Type to Function* mapping
• [IMPALA-4309] - Introduce expr rewrite phase.
• [IMPALA-4320] - Use gold linker by default
• [IMPALA-4439] - Kudu DML should report the number of rows that could not be modified
• [IMPALA-4450] - random query generator: prefer || for postgres string concatenation
• [IMPALA-4465] - Don't hold process wide lock while serializing Runtime Profile in GetRuntimeProfileStr()
• [IMPALA-4488] - Impala session times out prematurely - even if there is a running query
• [IMPALA-4516] - Don't hold process wide lock connection_to_sessions_map_lock_ while cancelling query
• [IMPALA-4529] - Inefficient use of try/catch in parser

Bug

• [IMPALA-1150] - Impala should limit the per-query core estimate to Yarn's per-node limit
• [IMPALA-1473] - Incorrect cardinality displayed in exec summary for merging exchange on coordinator.
• [IMPALA-1702] - "invalidate metadata" can cause duplicate TableIds
• [IMPALA-1862] - Invalid bool value not reported as a scanner error
• [IMPALA-2102] - cgroup not removed when cleaning up fragment after cancellation
• [IMPALA-2182] - MemTracker hierarchy may not correctly handle large RM limits
• [IMPALA-2799] - Query hang up if remote impalad hosts shut down
• [IMPALA-2864] - Ensure that client connections are closed after a failed RPC and a failed Open()
• [IMPALA-2967] - "Last Accessed Time" in Impalad's debug page is inconsistent with "Start Time"
• [IMPALA-3002] - Inconsistent accounting of exchange rows between impala-shell and profile summary
• [IMPALA-3125] - Incorrect assignment of outer join On-clause that only references one side of the join.
• [IMPALA-3126] - Incorrect assignment of an inner join On-clause predicate through an outer join.
• [IMPALA-3167] - Incorrect assignment of WHERE clause predicate through a grouping aggregation + outer join.
• [IMPALA-3178] - DROP DATABASE ... CASCADE is broken in Kudu
• [IMPALA-3299] - Stress test failures: Couldn't open transport for impala-stress-cdh5-trunk2-5.vpc.cloudera.com:22000
• [IMPALA-3314] - Altering table partition's storage format is not working and crashing the daemon
• [IMPALA-3342] - Runtime profile TotalCpuTime are often wrong
• [IMPALA-3346] - Kudu scanner : Improve perf of DecodeRowsIntoRowBatch
• [IMPALA-3347] - Kudu scanner : Expensive per row per column IsNull check
• [IMPALA-3348] - Kudu scanner : Avoid per slot check of Vector size in KuduScanner::RelocateValuesFromKudu
• [IMPALA-3349] - Kudu scanner : KuduScanBatch::RowPtr::Get* has very expensive checks in the hot path
• [IMPALA-3388] - Kudu should not throw table loading exceptions for fixable errors
• [IMPALA-3389] - Test hs2.test_hs2.TestHS2.test_get_log fails
• [IMPALA-3552] - Make incremental stats max serialized size configurable or dramatically increase limit.
• [IMPALA-3638] - Remove lazy creation of LLVM codegen module
• [IMPALA-3662] - Reduce parquet scanner memory usage
• [IMPALA-3671] - Add query option to limit scratch space usage
• [IMPALA-3710] - INSERT, UPDATE, DELETE should ignore conflicts by default
• [IMPALA-3713] - CRUD operations against Kudu tables should report the number of modified rows
• [IMPALA-3719] - Simplify CREATE table statements with Kudu tables
• [IMPALA-3724] - Support Kudu non-covering range partitions
• [IMPALA-3726] - Handle nullability and column encodings and compression for columns
• [IMPALA-3802] - test_create_alter_bulk_partition failure on Isilon: EOFException during ALTER TABLE
• [IMPALA-3815] - Clean up TopN node codegen'd code
• [IMPALA-3872] - Add support for caching PyPi packages
• [IMPALA-3874] - Predicates are not always pushed to Kudu
• [IMPALA-3875] - Thrift threaded server hang in some cases
• [IMPALA-3899] - Certain join types fail to spill after they start probing if there are matches stored in the hash table
• [IMPALA-3907] - Crash in test_udfs.py
• [IMPALA-3974] - IMPALA-2843 should consider using local_library_dir instead of java.io.tmpdir
• [IMPALA-4006] - bin/impala-config.sh contains dangerous rm -rf statements, which can delete everything it has access to
• [IMPALA-4016] - Avoid materializing slots only referenced by Kudu conjuncts
• [IMPALA-4020] - Catalog update can fail due to database creation/deletion in Hive.
• [IMPALA-4026] - Investigate regression introduced by "IMPALA-3629: Codegen TransferScratchTuples"
• [IMPALA-4028] - When the path to sentry-site.xml contains spaces, impalad and catalogd start up failed
• [IMPALA-4037] - ChildQuery::Cancel() appears to violate lock ordering
• [IMPALA-4038] - RPC delays for single query can lead to ImpalaServer not making progress on any queries
• [IMPALA-4042] - count(distinct NULL) fails on a view
• [IMPALA-4056] - toSql() of DistributeParam doesn't work as expected
• [IMPALA-4074] - Configuration items duplicate in template of YARN
• [IMPALA-4076] - Table with missing stats causes exception in RuntimeFilterGenerator.generateRuntimeFilters(RuntimeFilterGenerator.java:412)
• [IMPALA-4081] - Class rename will break old external data sources
• [IMPALA-4091] - Misconfigured backend unit tests do not generate logs.
• [IMPALA-4097] - Crash in kudu-scan-node-test
• [IMPALA-4098] - DCHECK in ExprContext::Clone() because the context has not been opened.
• [IMPALA-4102] - Remote Kudu reads should be reported
• [IMPALA-4104] - Fix TestEnv and reenable DCHECK in MemTracker::ConsumeLocal()
• [IMPALA-4108] - Move cached set_is_null/is_null/etc functions from SlotDescriptor to LlvmCodeGen
• [IMPALA-4111] - Backend death tests should not produce minidumps
• [IMPALA-4117] - Factor out test utility classes from simple-scheduler-test.cc
• [IMPALA-4120] - Incorrect results with LEAD() analytic function
• [IMPALA-4122] - qgen: cluster unit tests are bitrotted
• [IMPALA-4135] - Thrift threaded server times-out waiting connections during high load
• [IMPALA-4136] - testKudu planner test hangs if Kudu is not supported
• [IMPALA-4138] - RowBatch::AcquireState() cannot be used if the initial capacity of a RowBatch is not known
• [IMPALA-4153] - count(*) is not working properly on all blank('') columns in RCFile stored tables
• [IMPALA-4170] - Incorrect double quoting of identifier in SQL generated by COMPUTE INCREMENTAL STATS.
• [IMPALA-4180] - Crash: impala::DiskIoRequestContext::Cancel
• [IMPALA-4193] - Improve detection of maximum cpu clock frequency in cpu-info
• [IMPALA-4194] - Bump version to 2.8.0
• [IMPALA-4196] - Certain builtin functions throw symbol does not exist error when used inside aggregate function
• [IMPALA-4199] - Add 'SNAPSHOT' to unreleased versions in save-version.sh
• [IMPALA-4204] - be test failure: KuduScanNodeTest: TestLimitsAreEnforced: row mismatch
• [IMPALA-4205] - tmp-file-mgr-test: AddressSanitizer: heap-use-after-free
• [IMPALA-4206] - Complex view creation with subquery hits Preconditions check when computing column lineage.
• [IMPALA-4207] - data load failure: load_nested.py: AttributeError: 'Namespace' object has no attribute 'use_hive'
• [IMPALA-4213] - Planner not pushing some predicates with constant exprs to Kudu
• [IMPALA-4231] - Performance regression in TPC-H Q2 due to delay in filter arrival
• [IMPALA-4234] - Remove astylerc in be/
• [IMPALA-4237] - Incorrect results when materializing a decimal in a data source scan node.
• [IMPALA-4239] - BufferPoolTest.PinWithoutReservation and BufferPoolTest.CapacityExhausted fail on release build
• [IMPALA-4241] - "Child queries finished" message appears even when no child queries are present.
• [IMPALA-4246] - SleepForMs() utility function has undefined behavior for > 1s
• [IMPALA-4254] - impala-kudu unable to open lzo compressed tables
• [IMPALA-4260] - Alter table add column drops all the column stats
• [IMPALA-4262] - LZO-scanner fails when reading large index files from S3
• [IMPALA-4266] - Java udf expression returning string in group by can give incorrect results.
• [IMPALA-4274] - Jenkins builds can hang at the BE test "buffered-block-mgr-test"
• [IMPALA-4285] - Parquet scanner with MT_DOP > 0 crashes when materializing no slots.
• [IMPALA-4286] - DCHECK with MT_DOP=4 and query with nested types.
• [IMPALA-4296] - hadoop-next in gerrit has broken push_to_asf.py
• [IMPALA-4299] - run-all.sh command fails because HiveServer2 service failed to start
• [IMPALA-4301] - IGNORE NULLS clause in analytic function fails with WHERE-clause subqueries.
• [IMPALA-4302] - In-predicate filters do not scale as expected with # of elements constant IN-list
• [IMPALA-4303] - Incorrect change from UNION ALL to UNION DISTINCT after subquery rewriting.
• [IMPALA-4310] - push_to_asf.py doesn't respect --apache_remote
• [IMPALA-4322] - test_scanners_fuzz.py hits a DCHECK
• [IMPALA-4325] - StmtRewrite lost parentheses of CompoundPredicate in pushNegationToOperands()
• [IMPALA-4329] - Crash in scheduler when no backends are registered
• [IMPALA-4330] - Python failure in our builds
• [IMPALA-4332] - DCHECK in scheduler when querying Kudu or HBase table with MT_DOP > 0.
• [IMPALA-4333] - Crash in PlanRootSink::GetNext()
• [IMPALA-4335] - fetch calls may now produce empty row batches
• [IMPALA-4336] - Wrong results with nested union operands due to missing casts.
• [IMPALA-4339] - Coredumps don't end up in IMPALA_HOME if buildall.sh run from a different directory
• [IMPALA-4342] - "Cannot drop non-empty database: udf_test_errors" when using local filesystem
• [IMPALA-4348] - TestRPCTimeout.test_execplanfragment_timeout fails in exhaustive
• [IMPALA-4350] - Crash in PartitionedHashJoinNode::NodeDebugString() at log level 2
• [IMPALA-4357] - Fix DROP TABLE to pass analysis if the table fails to load
• [IMPALA-4362] - PFE ExecTime counter is always 0
• [IMPALA-4363] - SELECTing invalid timestamp value from Parquet file crashes impalad
• [IMPALA-4368] - SELECT NULL fails with MT_DOP > 0
• [IMPALA-4369] - DCHECK in Parquet scanner with MT_DOP > 1 when reading file with bad metadata.
• [IMPALA-4372] - 'Describe formatted' returns types in upper case
• [IMPALA-4374] - Change the stress test DDL statements for TPC-H and TPC-DS to using the new syntax for CREATE TABLE for Kudu
• [IMPALA-4377] - Failure in UdfExecutorTest.HiveStringsTest
• [IMPALA-4379] - Fix Kudu table type checking
• [IMPALA-4383] - Plan fragments may never send reports to coordinator
• [IMPALA-4384] - NPE in creating external Kudu table
• [IMPALA-4387] - Avro scanner crashes if the file schema has invalid decimal precision or scale
• [IMPALA-4388] - OOM in test_insert where none was before
• [IMPALA-4391] - HdfsScanners drops Status on the floor in multiple places
• [IMPALA-4392] - Add HdfsTableSink memory consumption to the query profile
• [IMPALA-4397] - Codegen for compute stats query on 1K column table takes 4 minutes
• [IMPALA-4408] - ASAN detected heap-buffer-overflow in Kudu scanner.
• [IMPALA-4409] - ImpalaServer::CancelInternal() can deadlock
• [IMPALA-4410] - PlanFragmentExecutor::Close() may crash if Prepare() fails
• [IMPALA-4411] - Kudu inserts violate lock ordering and could deadlock
• [IMPALA-4412] - Per operator timing in profile summary is incorrect when mt_dop > 0
• [IMPALA-4415] - DCHECK failure in SimpleScheduler::CreateScanInstances()
• [IMPALA-4421] - Rogue variable expansion creates "${RESULTS_DIR" directory inside tests/, which confuses run_tests.py
• [IMPALA-4423] - Wrong results with several conjunctive EXISTS subqueries that can be evaluated at query-compile time.
• [IMPALA-4425] - after merge IMPALA-4314: Standardize on MT-related data structures runtime Profile incorrect count
• [IMPALA-4432] - Cannot execute IR UDF when single node execution is enabled
• [IMPALA-4436] - StringValue::StringCompare bug for non-ascii string
• [IMPALA-4437] - Occasional crash in buffered-block-mgr-test
• [IMPALA-4440] - dcheck failure: lineage-util.h:91] Check failed: lineage.ended >= lineage.started
• [IMPALA-4441] - Divide-by-zero in RuntimeProfile::SummaryStatsCounter::SetStats
• [IMPALA-4442] - FE ParserTests may report UnsatisfiedLinkError
• [IMPALA-4444] - Use-after-free in memory returned from parquet scanner.
• [IMPALA-4446] - expr-test fails under ASAN
• [IMPALA-4452] - DCHECK in AggFnEvaluator::Init()
• [IMPALA-4453] - test_kudu_crud fails with wrong number of results
• [IMPALA-4454] - test_kudu.TestShowCreateTable fails on S3
• [IMPALA-4455] - MemPoolTest.TryAllocateAligned failed
• [IMPALA-4458] - DCHECK when running test_cancellation.py with MT_DOP > 0
• [IMPALA-4469] - Kudu tables created before IMPALA-3719 cannot load after upgrade
• [IMPALA-4470] - Predicate hits NumberFormatException due to constant folding.
• [IMPALA-4478] - Account for Kudu client memory in MemTracker
• [IMPALA-4479] - Evaluation of a constant boolean expr results in a NULL instead of FALSE.
• [IMPALA-4490] - Check failed: false Unexpected plan node with runtime filters
• [IMPALA-4493] - StringCompareTest fails under ASAN
• [IMPALA-4494] - Crash in SimpleScheduler when restarting under load
• [IMPALA-4495] - Runtime filters are disabled based on stats before they even arrive, contributing to performance cliff on TPC-H Q2
• [IMPALA-4497] - Kudu client usage on secure clusters results in crash
• [IMPALA-4498] - Timestamp timezone conversion causes a crash
• [IMPALA-4502] - test_partition_ddl_predicates breaks on non-HDFS filesystems
• [IMPALA-4504] - fragment-exec-state.cc:70] Check failed: status.ok() || done
• [IMPALA-4509] - Impala should provide SASL a mutex
• [IMPALA-4511] - Regression : Reported Fragment total time is incorrect
• [IMPALA-4514] - test_null_data failures in exhaustive runs
• [IMPALA-4517] - Stress test crash: impala::LlvmCodeGen::FinalizeModule
• [IMPALA-4518] - behavior with LEAD/LAG on string column/expression where offset goes beyond row bounds differs from postgres
• [IMPALA-4519] - TestFragmentLifecycle is flaky
• [IMPALA-4522] - Impalad crash under stress due to too many Kudu client threads
• [IMPALA-4523] - fix max varchar length
• [IMPALA-4525] - Codegen failure leads to query crash
• [IMPALA-4526] - Crash in test_mem_usage_scaling
• [IMPALA-4527] - Columns in Kudu tables created from Impala default to "NOT NULL"
• [IMPALA-4532] - Crashes in impala::ThreadCounterMeasurement::Stop()
• [IMPALA-4535] - ASAN build fails to compile
• [IMPALA-4539] - Parquet scanner memory bug: I/O buffer is attached to output batch while scratch batch rows still reference it
• [IMPALA-4540] - Scheduler calls function in DCHECK, leading to crash
• [IMPALA-4541] - test_codegen_mem_limit is flaky
• [IMPALA-4542] - Use-after-free in various backend tests
• [IMPALA-4550] - Mismatching types in JOIN crash Impala
• [IMPALA-4554] - Memory corruption of nested collection with MT_DOP > 0.
• [IMPALA-4557] - test_alloc_fail.py is flaky
• [IMPALA-4561] - Kudu syntax DISTRIBUTE BY should be PARTITIONED BY
• [IMPALA-4562] - impala crashes starting up on kerberized clusters where kudu is not supported
• [IMPALA-4564] - TestMtDopParquet.test_parquet should be skipped in legacy AGG/JOIN builds
• [IMPALA-4565] - Incorrect results under ASAN with mt_dop > 0 and old aggs/joins
• [IMPALA-4566] - Kudu client glog contention can cause timeouts
• [IMPALA-4567] - TestKuduOperations.test_kudu_alter_table fails
• [IMPALA-4569] - Multi-threaded compute stats DCHECK: impala::MemPool::~MemPool
• [IMPALA-4570] - Impala shell tarball is not usable on systems with setuptools versions where '0.7' is a substring of the full version string
• [IMPALA-4571] - InList predicates not being pushed to Kudu scans
• [IMPALA-4574] - The uuid() function should not be treated as a constant expr.
• [IMPALA-4578] - Inferred predicates not assigned to Kudu scan nodes.
• [IMPALA-4579] - SHOW CREATE VIEW fails for view containing WHERE clause on an unsecure cluster
• [IMPALA-4580] - Crash when FETCH_FIRST exhausts result set cache
• [IMPALA-4584] - test_kudu_alter_table fails on exhaustive
• [IMPALA-4585] - test_udfs.py fails on S3 and local filesystem builds
• [IMPALA-4586] - Values of non-deterministic UDFs are cached in backend
• [IMPALA-4589] - Kudu now requires replication factor must be odd; tests broken
• [IMPALA-4590] - Kudu test failure; several error messages changed
• [IMPALA-4592] - IllegalStateException when using nondeterministic functions in partition filter
• [IMPALA-4594] - WriteSlot and CodegenWriteSlot handle escaped NULL slots differently
• [IMPALA-4595] - Linking IR UDF module to main module crashes Impala
• [IMPALA-4608] - Fragment instance completion times can be very wrong
• [IMPALA-4609] - Make it clear that fragment thread counters are totals across all threads
• [IMPALA-4613] - Fragment total time regressed after IMPALA-4504
• [IMPALA-4614] - test_aggregation.py fails with IllegalStateException on Kudu table
• [IMPALA-4628] - test_kudu_alter_table fails with "Invalid alter step type"
• [IMPALA-4630] - Remove suspicious (but harmless) base64-encoded javascript in common-footer.tmpl
• [IMPALA-4633] - change broken kudu client default
• [IMPALA-4638] - Queries with mt_dop > 0 skip admission control
• [IMPALA-4642] - TestFragmentLifecycle broken; blocking gvos
• [IMPALA-4654] - Kudu scanner threads take a long time to close
• [IMPALA-4657] - "drop database cascade" command is not deleting KUDU tables
• [IMPALA-4662] - with Kudu tables, ClassCastException with NULL literal in IN list
• [IMPALA-4681] - buildall.sh -notests runs forever and hogs resources
• [IMPALA-4739] - ExprRewriter fails on HAVING clauses
• [IMPALA-4929] - Unsafe concurrent access to LlvmCodeGen::fn_refs_map_ can lead to unresolved symbols in LLVM

Task

• [IMPALA-3705] - Data loading for Kudu tables
• [IMPALA-3728] - Ensure partitioned tables do not allow mixed Kudu / non-Kudu partitions
• [IMPALA-3739] - Initial stress tests on Kudu
• [IMPALA-3771] - Expose kudu client timeouts or set to more conservative values
• [IMPALA-4109] - "Support" for Kudu in Impala on all platforms supported by Kudu
• [IMPALA-4112] - RQG-on-HIve: Remove IS [NOT] DISTINCT FROM clauses from Hive generated queries
• [IMPALA-4188] - run query generator with KUDU_IS_SUPPORTED=true
• [IMPALA-4283] - Ensure Kudu-specific DML audit & lineage behavior
• [IMPALA-4380] - Remove "cloudera" from bin/generate_minidump_collection_testdata.py
• [IMPALA-4427] - remove mention of cloudera/impala docker image from code base
• [IMPALA-4466] - Kudu DML functional tests
• [IMPALA-4467] - Kudu stress testing
• [IMPALA-4477] - Update Kudu toolchain build and bump client version

Sub-task

• [IMPALA-2848] - Simplify creation of a new Kudu table
• [IMPALA-3156] - Analyze the passed columns in DistributeComponent.java
• [IMPALA-3201] - Implement basic in-memory buffer pool
• [IMPALA-3204] - Integrate buffer pool reservations with memtrackers
• [IMPALA-3229] - Allow all contributors (or at least committers?) to run GVM
• [IMPALA-3567] - Factor out join build-side
• [IMPALA-3583] - Optimize HdfsTable::perPartitionFileDescMap_ to reduce memory usage
• [IMPALA-3653] - Consider using listLocatedStatus() API to get filestatus and blocklocations in one RPC call
• [IMPALA-3720] - Refactor CREATE TABLE grammar rules in prep for PRIMARY KEY
• [IMPALA-3721] - Simplify creating external Kudu tables and add DROP DATABASE CASCADE
• [IMPALA-3723] - Remove DDL delegate from the catalog for processing DDL statements on Kudu tables
• [IMPALA-3786] - Replace "cloudera" in code and package names with "impala"
• [IMPALA-4008] - Generate codegen module without operator/query-specific constants
• [IMPALA-4009] - document setting up oracle for use as qgen reference database
• [IMPALA-4023] - Don't attach BufferedTupleStreams to row batches
• [IMPALA-4146] - Basic load gen script for admission control
• [IMPALA-4171] - Remove testdata/udfs/impala-hive-udfs.jar
• [IMPALA-4176] - Add batched bit-unpacking utility functions
• [IMPALA-4270] - Planner should disallow queries with mt_dop > 0 that are not executable.
• [IMPALA-4338] - create kudu tables in postgres with primary keys
• [IMPALA-4340] - validate postgres 9.5 for random query generator reference database
• [IMPALA-4343] - qgen: model INSERT query
• [IMPALA-4352] - random query generator: add attributes to Table, Column for tracking of primary keys
• [IMPALA-4354] - qgen: implement SqlWriter methods for INSERT statements
• [IMPALA-4406] - Ensure Impala follows ASF crypto regulations
• [IMPALA-4572] - COMPUTE STATS on Parquet tables uses MT_DOP=4 by default
• [IMPALA-4577] - Adjust maximum size of row batch queue for non-Parquet scans with MT_DOP>0.