You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@doris.apache.org by "mrhhsg (via GitHub)" <gi...@apache.org> on 2026/04/03 02:11:14 UTC

[PR] [fix](be) Restore workload group limit refresh after memory recovery [doris]

mrhhsg opened a new pull request, #62070:
URL: https://github.com/apache/doris/pull/62070

   ### What problem does this PR solve?
   
   Issue Number: close #xxx
   
   Related PR: #xxx
   
   Problem Summary:
   
   ### Release note
   
   None
   
   ### Check List (For Author)
   
   - Test <!-- At least one of them must be included. -->
       - [ ] Regression test
       - [ ] Unit Test
       - [ ] Manual test (add detailed scripts or steps below)
       - [ ] No need to test or manual test. Explain why:
           - [ ] This is a refactor/code format and no logic has been changed.
           - [ ] Previous test can cover this change.
           - [ ] No code files have been changed.
           - [ ] Other reason <!-- Add your reason?  -->
   
   - Behavior changed:
       - [ ] No.
       - [ ] Yes. <!-- Explain the behavior change -->
   
   - Does this need documentation?
       - [ ] No.
       - [ ] Yes. <!-- Add document PR link here. eg: https://github.com/apache/doris-website/pull/1214 -->
   
   ### Check List (For Reviewer who merge this PR)
   
   - [ ] Confirm the release note
   - [ ] Confirm test cases
   - [ ] Confirm document
   - [ ] Add branch pick label <!-- Add branch pick label that this PR should merge into -->
   
   


-- 
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: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [fix](be) Restore workload group limit refresh after memory recovery [doris]

Posted by "mrhhsg (via GitHub)" <gi...@apache.org>.
mrhhsg commented on PR #62070:
URL: https://github.com/apache/doris/pull/62070#issuecomment-4181583401

   run buildall


-- 
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: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [fix](be) Restore workload group limit refresh after memory recovery [doris]

Posted by "hello-stephen (via GitHub)" <gi...@apache.org>.
hello-stephen commented on PR #62070:
URL: https://github.com/apache/doris/pull/62070#issuecomment-4181833213

   # BE UT Coverage Report
   Increment line coverage ` ` :tada:
   
   [Increment coverage report](http://coverage.selectdb-in.cc/coverage/bba902b69dda463b4b6358fc58df47884a91dca3_bba902b69dda463b4b6358fc58df47884a91dca3/increment_report/index.html)
   [Complete coverage report](http://coverage.selectdb-in.cc/coverage/bba902b69dda463b4b6358fc58df47884a91dca3_bba902b69dda463b4b6358fc58df47884a91dca3/report/index.html)
   | Category          | Coverage           |
   |-------------------|--------------------|
   | Function Coverage | 52.95% (20061/37885) |
   | Line Coverage     | 36.58% (188563/515539) |
   | Region Coverage   | 32.83% (146410/445901) |
   | Branch Coverage   | 33.99% (64119/188625) |


-- 
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: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [fix](be) Restore workload group limit refresh after memory recovery [doris]

Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #62070:
URL: https://github.com/apache/doris/pull/62070#issuecomment-4181715798

   
   <details>
   <summary>TPC-H: <b>Total hot run time: 29204 ms</b></summary>
   
   ```
   machine: 'aliyun_ecs.c7a.8xlarge_32C64G'
   scripts: https://github.com/apache/doris/tree/master/tools/tpch-tools
   Tpch sf100 test result on commit bba902b69dda463b4b6358fc58df47884a91dca3, data reload: false
   
   ------ Round 1 ----------------------------------
   orders	Doris	NULL	NULL	0	0	0	NULL	0	NULL	NULL	2023-12-26 18:27:23	2023-12-26 18:42:55	NULL	utf-8	NULL	NULL	
   ============================================
   q1	17910	3787	3759	3759
   q2	q3	10680	924	611	611
   q4	4688	466	355	355
   q5	7466	1358	1164	1164
   q6	183	162	136	136
   q7	923	946	777	777
   q8	9310	1386	1373	1373
   q9	5458	5330	5343	5330
   q10	6246	2033	1806	1806
   q11	462	272	277	272
   q12	647	411	288	288
   q13	18119	2796	2151	2151
   q14	286	293	258	258
   q15	q16	862	858	788	788
   q17	1023	1046	809	809
   q18	6447	5691	5551	5551
   q19	1181	1273	1031	1031
   q20	552	421	301	301
   q21	4892	2544	2085	2085
   q22	506	408	359	359
   Total cold run time: 97841 ms
   Total hot run time: 29204 ms
   
   ----- Round 2, with runtime_filter_mode=off -----
   orders	Doris	NULL	NULL	150000000	42	6422171781	NULL	22778155	NULL	NULL	2023-12-26 18:27:23	2023-12-26 18:42:55	NULL	utf-8	NULL	NULL	
   ============================================
   q1	4559	4498	4681	4498
   q2	q3	4630	4714	4246	4246
   q4	2022	2075	1348	1348
   q5	4899	5053	5189	5053
   q6	194	163	126	126
   q7	2014	1732	1636	1636
   q8	3439	3148	3206	3148
   q9	8283	8351	8345	8345
   q10	4529	4592	4245	4245
   q11	608	435	393	393
   q12	668	714	495	495
   q13	2655	3034	2304	2304
   q14	291	298	277	277
   q15	q16	771	787	694	694
   q17	1305	1291	1236	1236
   q18	7898	7315	6991	6991
   q19	1142	1141	1120	1120
   q20	2277	2282	1949	1949
   q21	6032	5301	4766	4766
   q22	536	508	420	420
   Total cold run time: 58752 ms
   Total hot run time: 53290 ms
   ```
   </details>
   


-- 
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: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [fix](be) Restore workload group limit refresh after memory recovery [doris]

Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on PR #62070:
URL: https://github.com/apache/doris/pull/62070#issuecomment-4181517486

   No issues found in this review.
   
   Checkpoint conclusions:
   - Goal/task: The patch addresses the reported early-return limit-refresh bug and the paused-query cleanup/recovery branches. The added unit tests cover the zero-WG-memory refresh path, Phase 2/3/4 paused-query handling, and cross-WG blocking scenarios.
   - Scope/minimality: The change stays focused to `be/src/runtime/workload_group/workload_group_manager.cpp` plus targeted BE unit tests.
   - Concurrency: The touched shared state remains guarded by the existing `_paused_queries_lock`, and the changed `WorkloadGroup` accesses continue to use the existing WG locking model. I did not find a new lock-order or deadlock regression in the modified paths.
   - Lifecycle/static init/config/compatibility: No new lifecycle hazards, static initialization issues, config items, or compatibility concerns were introduced.
   - Parallel paths/conditions: The fix is applied consistently across the affected paused-query branches, and the new same-WG/global recent-cancel conditions match the surrounding memory semantics.
   - Test coverage: Coverage is materially improved for the fixed branches. I could not rerun `./run-be-ut.sh --run --filter="WorkloadGroupManagerTest.*"` in this runner because `thirdparty/installed/bin/protoc` is missing, so execution here could not validate the tests.
   - Observability: Existing INFO logging remains sufficient for these branches.
   - Transaction/persistence/data-write/FE-BE protocol: Not applicable to this change.
   - Performance/other issues: No additional correctness or performance issues were identified in the touched logic.
   


-- 
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: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [fix](be) Restore workload group limit refresh after memory recovery [doris]

Posted by "Thearas (via GitHub)" <gi...@apache.org>.
Thearas commented on PR #62070:
URL: https://github.com/apache/doris/pull/62070#issuecomment-4181412252

   
   Thank you for your contribution to Apache Doris.
   Don't know what should be done next? See [How to process your PR](https://cwiki.apache.org/confluence/display/DORIS/How+to+process+your+PR).
   
   Please clearly describe your PR:
   1. What problem was fixed (it's best to include specific error reporting information). How it was fixed.
   2. Which behaviors were modified. What was the previous behavior, what is it now, why was it modified, and what possible impacts might there be.
   3. What features were added. Why was this function added?
   4. Which code was refactored and why was this part of the code refactored?
   5. Which functions were optimized and what is the difference before and after the optimization?
       


-- 
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: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [fix](be) Restore workload group limit refresh after memory recovery [doris]

Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on PR #62070:
URL: https://github.com/apache/doris/pull/62070#issuecomment-4181516852

   No issues found in this review.
   
   Checkpoint conclusions:
   - Goal/task: The patch addresses the reported early-return limit-refresh bug and the paused-query cleanup/recovery branches. The added unit tests cover the zero-WG-memory refresh path, Phase 2/3/4 paused-query handling, and cross-WG blocking scenarios.
   - Scope/minimality: The change stays focused to  plus targeted BE unit tests.
   - Concurrency: The touched shared state remains guarded by the existing , and the changed  accesses continue to use the existing WG locking model. I did not find a new lock-order or deadlock regression in the modified paths.
   - Lifecycle/static init/config/compatibility: No new lifecycle hazards, static initialization issues, config items, or compatibility concerns were introduced.
   - Parallel paths/conditions: The fix is applied consistently across the affected paused-query branches, and the new same-WG/global recent-cancel conditions match the surrounding memory semantics.
   - Test coverage: Coverage is materially improved for the fixed branches. I could not rerun Target system: Linux; Target arch: x86_64
   Python 3.12.3
   Check JAVA_HOME version
   Apache Maven 3.9.14 (996c630dbc656c76214ce58821dcc58be960875b)
   Maven home: /usr/share/apache-maven-3.9.14
   Java version: 17.0.18, vendor: Eclipse Adoptium, runtime: /usr/lib/jvm/temurin-17-jdk-amd64
   Default locale: en_US, platform encoding: UTF-8
   OS name: "linux", version: "6.17.0-1008-azure", arch: "amd64", family: "unix"
   cmake version 3.31.6
   
   CMake suite maintained and supported by Kitware (kitware.com/cmake).
   ninja 1.13.2
   Get params:
       PARALLEL            -- 1
       CLEAN               -- 0
       ENABLE_PCH          -- ON
       WITH_TDE_DIR        -- 
   
   Build Backend UT
   Update apache-orc submodule ...
   Update clucene submodule ...
   -- Make program: /usr/local/bin/ninja
   -- Use ccache:  and 
   -- Extra cxx flags: 
   -- GLIBC_COMPATIBILITY is ON
   -- USE_LIBCPP is OFF
   -- USE_JEMALLOC is OFF
   -- USE_UNWIND is ON
   -- ENABLE_PCH is ON
   -- USE_AVX2 is ON
   -- Build type is ASAN_UT
   -- Build target arch is x86_64
   -- DORIS_HOME is /home/runner/work/doris/doris
   -- THIRDPARTY_DIR is /home/runner/work/doris/doris/thirdparty/installed
   -- make test: ON
   -- make benchmark: OFF
   -- build fs benchmark tool: OFF
   -- build task executor simulator: OFF
   -- build file cache lru tool: OFF
   -- build gensrc if necessary
   make: Entering directory '/home/runner/work/doris/doris/gensrc'
   make -C script
   make[1]: Entering directory '/home/runner/work/doris/doris/gensrc/script'
   /home/runner/work/doris/doris/gensrc/script/gen_build_version.sh
   make[1]: Leaving directory '/home/runner/work/doris/doris/gensrc/script'
   make -C proto
   make[1]: Entering directory '/home/runner/work/doris/doris/gensrc/proto'
   /home/runner/work/doris/doris/thirdparty/installed/bin/protoc --proto_path=/home/runner/work/doris/doris/gensrc/proto --cpp_out=/home/runner/work/doris/doris/gensrc/proto/../build//gen_cpp /home/runner/work/doris/doris/gensrc/proto/descriptors.proto
   make[1]: Leaving directory '/home/runner/work/doris/doris/gensrc/proto'
   make: Leaving directory '/home/runner/work/doris/doris/gensrc'
   -- Configuring incomplete, errors occurred! in this runner because  is missing, so execution here could not validate the tests.
   - Observability: Existing INFO logging remains sufficient for these branches.
   - Transaction/persistence/data-write/FE-BE protocol: Not applicable to this change.
   - Performance/other issues: No additional correctness or performance issues were identified in the touched logic.


-- 
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: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [fix](be) Restore workload group limit refresh after memory recovery [doris]

Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on PR #62070:
URL: https://github.com/apache/doris/pull/62070#issuecomment-4213255501

   PR approved by anyone and no changes requested.


-- 
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: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [fix](be) Restore workload group limit refresh after memory recovery [doris]

Posted by "mrhhsg (via GitHub)" <gi...@apache.org>.
mrhhsg commented on PR #62070:
URL: https://github.com/apache/doris/pull/62070#issuecomment-4181501752

   /review


-- 
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: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [fix](be) Restore workload group limit refresh after memory recovery [doris]

Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #62070:
URL: https://github.com/apache/doris/pull/62070#issuecomment-4181768771

   
   <details>
   <summary>TPC-DS: <b>Total hot run time: 178838 ms</b></summary>
   
   ```
   machine: 'aliyun_ecs.c7a.8xlarge_32C64G'
   scripts: https://github.com/apache/doris/tree/master/tools/tpcds-tools
   TPC-DS sf100 test result on commit bba902b69dda463b4b6358fc58df47884a91dca3, data reload: false
   
   query5	4363	706	528	528
   query6	356	233	216	216
   query7	4258	622	336	336
   query8	353	257	236	236
   query9	8734	3898	3896	3896
   query10	473	364	329	329
   query11	6652	5484	5135	5135
   query12	188	135	127	127
   query13	1298	605	448	448
   query14	5665	5149	4767	4767
   query14_1	4143	4152	4086	4086
   query15	209	205	185	185
   query16	994	464	445	445
   query17	1074	787	664	664
   query18	2448	503	379	379
   query19	248	213	176	176
   query20	137	134	129	129
   query21	222	141	119	119
   query22	13919	14938	14765	14765
   query23	18098	17173	16710	16710
   query23_1	16769	16837	17049	16837
   query24	7510	1741	1350	1350
   query24_1	1352	1373	1411	1373
   query25	605	485	426	426
   query26	1268	325	170	170
   query27	2693	612	380	380
   query28	4442	1893	1868	1868
   query29	971	660	561	561
   query30	295	229	193	193
   query31	1088	1052	940	940
   query32	87	74	70	70
   query33	539	364	279	279
   query34	1209	1153	670	670
   query35	728	789	666	666
   query36	1256	1224	1063	1063
   query37	151	99	88	88
   query38	3093	2974	2965	2965
   query39	909	890	848	848
   query39_1	839	852	822	822
   query40	242	148	129	129
   query41	62	63	58	58
   query42	112	105	108	105
   query43	310	318	276	276
   query44	
   query45	223	195	187	187
   query46	1136	1312	773	773
   query47	2338	2324	2233	2233
   query48	406	416	302	302
   query49	631	514	425	425
   query50	767	285	218	218
   query51	4346	4334	4451	4334
   query52	108	110	98	98
   query53	241	270	203	203
   query54	316	280	255	255
   query55	104	97	90	90
   query56	318	316	304	304
   query57	1760	1722	1649	1649
   query58	302	279	273	273
   query59	2916	2961	2711	2711
   query60	326	331	351	331
   query61	161	146	155	146
   query62	698	623	570	570
   query63	236	199	190	190
   query64	5313	1315	955	955
   query65	
   query66	1489	473	373	373
   query67	24441	24275	24171	24171
   query68	
   query69	459	339	308	308
   query70	1009	982	998	982
   query71	309	281	275	275
   query72	2986	2711	2420	2420
   query73	855	831	448	448
   query74	9871	9712	9583	9583
   query75	2739	2631	2310	2310
   query76	2315	1113	755	755
   query77	390	400	322	322
   query78	11430	11388	10780	10780
   query79	1527	1105	829	829
   query80	688	567	508	508
   query81	446	277	237	237
   query82	1318	169	125	125
   query83	366	301	277	277
   query84	281	155	121	121
   query85	854	545	438	438
   query86	379	327	312	312
   query87	3296	3208	3122	3122
   query88	3536	2686	2690	2686
   query89	444	384	359	359
   query90	1987	182	173	173
   query91	175	168	149	149
   query92	76	75	73	73
   query93	975	960	556	556
   query94	531	343	301	301
   query95	639	457	343	343
   query96	1069	774	336	336
   query97	2689	2683	2573	2573
   query98	242	226	224	224
   query99	1064	1074	981	981
   Total cold run time: 257101 ms
   Total hot run time: 178838 ms
   ```
   </details>
   


-- 
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: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [fix](be) Restore workload group limit refresh after memory recovery [doris]

Posted by "mrhhsg (via GitHub)" <gi...@apache.org>.
mrhhsg merged PR #62070:
URL: https://github.com/apache/doris/pull/62070


-- 
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: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [fix](be) Restore workload group limit refresh after memory recovery [doris]

Posted by "hello-stephen (via GitHub)" <gi...@apache.org>.
hello-stephen commented on PR #62070:
URL: https://github.com/apache/doris/pull/62070#issuecomment-4182007519

   # BE Regression && UT Coverage Report
   Increment line coverage `92.44% (428/463)` :tada:
   
   [Increment coverage report](http://coverage.selectdb-in.cc/coverage/62070_bba902b69dda463b4b6358fc58df47884a91dca3_merge/increment_report/index.html)
   [Complete coverage report](http://coverage.selectdb-in.cc/coverage/62070_bba902b69dda463b4b6358fc58df47884a91dca3_merge/report/index.html)
   | Category          | Coverage           |
   |-------------------|--------------------|
   | Function Coverage | 73.66% (27330/37105) |
   | Line Coverage     | 57.30% (294481/513969) |
   | Region Coverage   | 54.36% (244635/450040) |
   | Branch Coverage   | 56.17% (106273/189207) |


-- 
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: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [fix](be) Restore workload group limit refresh after memory recovery [doris]

Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on PR #62070:
URL: https://github.com/apache/doris/pull/62070#issuecomment-4213255283

   PR approved by at least one committer and no changes requested.


-- 
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: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org