Commit Graph

2648 Commits

Author SHA1 Message Date
Wangda Tan
b7384a8d02 YARN-10531. Be able to disable user limit factor for CapacityScheduler Leaf Queue. (Qi Zhu via wangda)
Change-Id: I670e5525619b320745254609c48e7e1afb084835
2021-01-26 08:24:28 -08:00
Szilard Nemeth
f1766e5bb4 YARN-10596. Allow static definition of childless ParentQueues with auto-queue-creation-v2 enabled. Contributed by Andras Gyori 2021-01-26 16:22:35 +01:00
Szilard Nemeth
8620984b8d YARN-10515. Fix flaky test TestCapacitySchedulerAutoQueueCreation.testDynamicAutoQueueCreationWithTags. Contributed by Peter Bacsko 2021-01-21 20:58:57 +01:00
Peter Bacsko
17fdac8de5 YARN-10581. CS Flexible Auto Queue Creation: Modify RM /scheduler endpoint to include queue creation type for queues. Contributed by Szilard Nemeth. 2021-01-21 18:06:53 +01:00
Szilard Nemeth
06fef5ee43 YARN-10579. CS Flexible AQC: Modify RM /scheduler endpoint to include weight values for queues. Contributed by Szilard Nemeth 2021-01-21 09:23:11 +01:00
bibinchundatt
8bc2dfbf36 YARN-10519. Refactor QueueMetricsForCustomResources class to move to yarn-common package. Contributed by Minni Mittal 2021-01-21 11:23:30 +05:30
Szilard Nemeth
cfe6e1f7da YARN-10578. Fix Auto Queue Creation parent handling. Contributed by Andras Gyori 2021-01-20 15:22:44 +01:00
Sunil G
4b5bc05a78 YARN-10512. CS Flexible Auto Queue Creation: Modify RM /scheduler endpoint to include mode of operation for CS. Contributed by Szilard Nemeth. 2021-01-20 12:28:59 +05:30
Szilard Nemeth
a326f22606 YARN-10573. Enhance placement rule conversion in fs2cs in weight mode and enable it by default. Contributed by Peter Bacsko 2021-01-19 09:42:40 +01:00
Szilard Nemeth
6abdb148e4 YARN-10535. Make queue placement in CapacityScheduler compliant with auto-queue-placement. Contributed by Gergely Pollak 2021-01-18 20:19:36 +01:00
Szilard Nemeth
acea862d59 YARN-10574. Fix the FindBugs warning introduced in YARN-10506. Contributed by Gergely Pollak 2021-01-18 16:04:45 +01:00
Wangda Tan
3d46141583 YARN-10506. Update queue creation logic to use weight mode and allow the flexible static/dynamic creation. (Contributed by Qi Zhu, Andras Gyori)
Change-Id: I118862fd5e11ee6888275e2bcf667fedfa56c5d7
2021-01-15 14:20:08 -08:00
Szilard Nemeth
836c630430 YARN-10525. Add weight mode conversion to fs2cs. Contributed by Peter Bacsko 2021-01-13 14:24:12 +01:00
Szilard Nemeth
85b1c017ee YARN-10563. Fix dependency exclusion problem in poms. Contributed by Peter Bacsko 2021-01-12 15:29:20 +01:00
Szilard Nemeth
9c6e093783 YARN-10570. Remove experimental warning message from fs2cs. Contributed by Peter Bacsko 2021-01-12 15:21:27 +01:00
Wangda Tan
b0eec09097 YARN-10504. Implement weight mode in Capacity Scheduler. (Contributed by Wangda Tan, Benjamin Teke, zhuqi, Andras Gyori)
Change-Id: Ic49c730b0ab502ba86527fb662d25c4c8b1c2588
2021-01-11 17:46:09 -08:00
Szilard Nemeth
4176759909 YARN-10507. Add the capability to fs2cs to write the converted placement rules inside capacity-scheduler.xml. Contributed by Peter Bacsko 2021-01-08 13:40:33 +01:00
Szilard Nemeth
580a6a75a3 YARN-10528. maxAMShare should only be accepted for leaf queues, not parent queues. Contributed by Siddharth Ahuja 2021-01-08 12:38:43 +01:00
srinivasst
1b1791075a
YARN-10538: Add RECOMMISSIONING nodes to the list of updated nodes returned to the AM (#2564)
Contributed by Srinivas S T
2021-01-08 10:52:52 +05:30
Neil
fa4cf91b57
YARN-10541. capture the performance metrics of ZKRMStateStore (#2568) 2020-12-30 22:20:12 -08:00
Akira Ajisaka
df7f1e5199
YARN-10499. TestRouterWebServiceREST fails (#2490). Contributed by Akira Ajisaka 2020-12-16 17:26:15 +01:00
Szilard Nemeth
df35c7f519 YARN-10526. RMAppManager CS Placement ignores parent path. Contributed by Gergely Pollak 2020-12-15 17:30:14 +01:00
zhuqi
d67ccd03e3
YARN-10380: Import logic of multi-node allocation in CapacityScheduler (#2494)
Contributed by Qi Zhu.
2020-12-09 19:48:39 +08:00
adamantal
9969745343 YARN-9883. Reshape SchedulerHealth class. Contributed by D M Murali Krishna Reddy 2020-12-03 09:55:06 +01:00
Eric Payne
fa773a8326 YARN-10278: CapacityScheduler test framework ProportionalCapacityPreemptionPolicyMockFramework. Contributed by Szilard Nemeth (snemeth) 2020-12-01 22:06:47 +00:00
Ankit Kumar
08b2e285db
YARN-10488. Several typos in package: org.apache.hadoop.yarn.server.resourcemanager.scheduler.fair (#2485)
Co-authored-by: akumar <akumar@cloudera.com>
Reviewed-by: Wei-Chiu Chuang <weichiu@apache.org>
Reviewed-by: Akira Ajisaka <aajisaka@apache.org>
2020-11-25 13:00:09 +09:00
Szilard Nemeth
0d3155a687 YARN-10457. Add a configuration switch to change between legacy and JSON placement rule format. Contributed by Gergely Pollak 2020-11-18 17:02:47 +01:00
Szilard Nemeth
5ff70a59c4 YARN-10486. FS-CS converter: handle case when weight=0 and allow more lenient capacity checks in Capacity Scheduler. Contributed by Peter Bacsko 2020-11-18 16:51:42 +01:00
Peter Bacsko
567600fd80 YARN-10425. Replace the legacy placement engine in CS with the new one. Contributed by Gergely Pollak. 2020-11-11 17:10:15 +01:00
Peter Bacsko
cd0490e8c6 Revert "YARN-10425. Replace the legacy placement engine in CS with the new one. Contributed by Gergely Pollak."
This reverts commit b0ab222a6c.
2020-11-11 17:06:12 +01:00
Peter Bacsko
b0ab222a6c YARN-10425. Replace the legacy placement engine in CS with the new one. Contributed by Gergely Pollak. 2020-11-11 17:03:16 +01:00
Eric E Payne
0461a07c01 YARN-10475: Scale RM-NM heartbeat interval based on node utilization. Contributed by Jim Brennan (Jim_Brennan). 2020-11-02 16:55:06 +00:00
Surendra Singh Lilhore
7169ec4509
YARN-10442. RM should make sure node label file highly available. (#2390)
* YARN-10442. RM should make sure node label file highly available.

* YARN-10442. Corrected method name.
            Added license header for file TestNodeLabelFileReplication.java
            Added Test code.

* YARN-10442. Changed property name and log.

* YARN-10442. Changed default replication for FS store file.
2020-10-30 23:05:16 +05:30
Wangda Tan
c47c9fd65d YARN-10458. Hive On Tez queries fails upon submission to dynamically created pools. (Peter Bacsko via wangda)
Change-Id: I518dc925187ce55e9d35a37ba20878c0f4e37e5c
2020-10-30 09:49:48 -07:00
Jonathan Hung
bab5bf9743 YARN-10467. ContainerIdPBImpl objects can be leaked in RMNodeImpl.completedContainers. Contributed by Haibo Chen 2020-10-28 10:33:18 -07:00
Sunil G
7435604a91 YARN-10453. Add partition resource info to get-node-labels and label-mappings api responses. Contributed Akhil PB. 2020-10-22 12:52:58 +05:30
Ayush Saxena
1e3a6efcef
HADOOP-17288. Use shaded guava from thirdparty. (#2342). Contributed by Ayush Saxena. 2020-10-17 12:01:18 +05:30
Eric Badger
8b8c672780 YARN-10450. Add cpu and memory utilization per node and cluster-wide metrics.
Contributed by Jim Brennan.
2020-10-15 17:57:43 +00:00
Szilard Nemeth
42358169f6 YARN-10454: Add applicationName policy. Contributed by Peter Bacsko 2020-10-13 10:11:10 +02:00
Jim Brennan
b361f29dda YARN-10451. RM (v1) UI NodesPage can NPE when yarn.io/gpu resource type is defined. Contributed by Eric Payne 2020-10-06 18:18:08 +00:00
Fei Hui
d0d10f7e8f
HADOOP-17276. Extend CallerContext to make it include many items (#2327) 2020-10-04 02:02:28 +09:00
Adam Antal
bb8446e80c YARN-10447. TestLeafQueue: ActivitiesManager thread might interfere with ongoing stubbing. Contributed by Peter Bacsko 2020-10-01 11:42:04 +02:00
Szilard Nemeth
a7a1f1541a YARN-10413. Change fs2cs to generate mapping rules in the new format. Contributed by Peter Bacsko 2020-09-29 20:18:26 +02:00
Shubham Gupta
e3cd627069
YARN-10438. Handle null containerId in ClientRMService#getContainerReport() (#2313)
Co-authored-by: Shubham Gupta <gshubham@microsoft.com>
2020-09-25 12:20:39 +05:30
Szilard Nemeth
eacbe07b56 YARN-9333. TestFairSchedulerPreemption.testRelaxLocalityPreemptionWithNoLessAMInRemainingNodes fails intermittently. Contributed by Peter Bacsko 2020-09-17 19:57:19 +02:00
Eric E Payne
9afec2ed17 YARN-10390: LeafQueue: retain user limits cache across assignContainers() calls. Contributed by Samir Khan (samkhan). 2020-09-11 13:29:26 +00:00
Szilard Nemeth
773ac799c6 YARN-10387. Implement logic which returns MappingRule objects based on mapping rules. Contributed by Peter Bacsko 2020-09-09 15:12:58 +02:00
Brahma Reddy Battula
43572fc7f8 YARN-10397. SchedulerRequest should be forwarded to scheduler if custom scheduler supports placement constraints. Contributed by Bilwa S T. 2020-09-09 17:08:13 +05:30
Peter Bacsko
aba4a506d6 YARN-10424. Adapt existing AppName and UserGroupMapping unittests to ensure backwards compatibility. Contributed by Benjamin Teke. 2020-09-09 09:41:11 +02:00
Peter Bacsko
c4fb4044b2 YARN-10415. Create a group matcher which checks ALL groups of the user. Contributed by Gergely Pollak. 2020-09-08 10:57:00 +02:00
Adam Antal
34fe74da0e YARN-10332. RESOURCE_UPDATE event was repeatedly registered in DECOMMISSIONING state. Contributed by yehuanhuan 2020-09-07 11:44:09 +02:00
Peter Bacsko
360bbcd3bc YARN-10411. Create an allowCreate flag for MappingRuleAction. Contributed by Gergely Pollak. 2020-09-07 11:39:03 +02:00
Peter Bacsko
1841a5bb03 YARN-10376. Create a class that covers the functionality of UserGroupMappingPlacementRule and AppNameMappingPlacementRule using the new mapping rules. Contributed by Gergely Pollak. 2020-09-05 10:00:33 +02:00
Szilard Nemeth
696e4fe50e Typos in YarnScheduler#allocate method's doc comment. Contributed by Siddharth Ahuja 2020-09-04 09:59:54 +02:00
Peter Bacsko
5c15815773 YARN-10375. CS Mapping rule config parser should return MappingRule objects. Contributed by Gergely Pollak. 2020-09-03 18:54:22 +02:00
Peter Bacsko
288dab7563 YARN-10372. Create MappingRule class to represent each CS mapping rule. Contributed by Gergely Pollak 2020-09-03 15:02:48 +02:00
Peter Bacsko
c7dab2b23e YARN-10374. Create Actions for CS mapping rules. Contributed by Gergely Pollak. 2020-09-01 17:38:49 +02:00
Adam Antal
73a0d149e2 YARN-10386. Addendum patch: Fixing ASF Warnings. Contributed by Peter Bacsko. 2020-09-01 14:39:24 +02:00
Szilard Nemeth
4ffe26f9b8 [UI1] Provide a way to hide Tools section in Web UIv1. Contributed by Andras Gyori 2020-08-31 16:00:39 +02:00
Szilard Nemeth
56a5c360a1 YARN-10373. Create Matchers for CS mapping rules. Contributed by Gergely Pollak 2020-08-29 21:34:55 +02:00
Szilard Nemeth
f4f872b778 YARN-10371. Create variable context class for CS queue mapping rules. Contributed by Gergely Pollak 2020-08-29 17:31:48 +02:00
Szilard Nemeth
a888d580d8 YARN-10409. Improve MockQueueHierarchyBuilder to detect queue ambiguity. Contributed by Gergely Pollak 2020-08-29 15:24:50 +02:00
Szilard Nemeth
44542863f4 YARN-10408. Extract MockQueueHierarchyBuilder to a separate class. Contributed by Gergely Pollak 2020-08-28 17:35:45 +02:00
Szilard Nemeth
6e3d7050a9 YARN-10386. Create new JSON schema for Placement Rules. Contributed by Peter Bacsko and Benjamin Teke 2020-08-28 17:09:31 +02:00
Prabhu Joseph
5c927eb550 YARN-10360. Support Multi Node Placement in SingleConstraintAppPlacementAllocator
Reviewed by Sunil G.
2020-08-24 15:00:02 +05:30
Zhengbo Li
9b9f7ea16a
YARN-10399 Refactor NodeQueueLoadMonitor class to make it extendable (#2228)
Refactor NodeQueueLoadMonitor class to make it extendable
2020-08-19 10:14:50 -07:00
Sunil G
82ec28f442 YARN-10396. Max applications calculation per queue disregards queue level settings in absolute mode. Contributed by Benjamin Teke. 2020-08-19 11:54:48 +05:30
Prabhu Joseph
6c2ce3d56b YARN-10389. Option to override RMWebServices with custom WebService class
Contributed by Tanu Ajmera. Reviewed by Bilwa ST and Sunil G.
2020-08-11 15:46:47 +05:30
Prabhu Joseph
5e0f879779 YARN-10364. Fix logic of queue capacity is absolute resource or percentage.
Contributed by Bilwa ST. Reviewed by Sunil G.
2020-08-08 22:33:58 +05:30
Jonathan Hung
64753addba YARN-10251. Show extended resources on legacy RM UI. Contributed by Eric Payne 2020-08-07 17:43:01 -07:00
bibinchundatt
81da221c75 YARN-10388. RMNode updatedCapability flag not set while RecommissionNodeTransition. Contributed by Pranjal Protim Borah 2020-08-07 08:36:52 +05:30
Prabhu Joseph
c7e71a6c0b YARN-10361. Make custom DAO classes configurable into RMWebApp#JAXBContextResolver.
Contributed by Bilwa ST.
2020-08-06 13:19:52 +05:30
Eric Badger
5edd8b925e YARN-4575. ApplicationResourceUsageReport should return ALL reserved resource.
Contributed by Bibin Chundatt and Eric Payne.
2020-08-05 18:47:49 +00:00
Prabhu Joseph
0277856738 YARN-10377. Fix filter index to show apps while clicking on queue in RM UI
Contributed by Tarun Parimi.
2020-08-04 23:10:45 +05:30
Prabhu Joseph
aa5afa72c5 YARN-10381. Add application attempt state in AppAttempts RM REST API
Contributed by Siddharth Ahuja. Reviewed by Bilwa ST.
2020-08-04 23:07:28 +05:30
bibinchundatt
5d8600e80a YARN-10369. Make NMTokenSecretManagerInRM sending NMToken for nodeId DEBUG. Contributed by Jim Brennan. 2020-07-29 23:26:08 +05:30
Jonathan Hung
3eaf62726f YARN-10343. Legacy RM UI should include labeled metrics for allocated, total, and reserved resources. Contributed by Eric Payne 2020-07-28 13:43:19 -07:00
bibinchundatt
5dadf963d3 YARN-10208. Add capacityScheduler metric for NODE_UPDATE interval. Contributed by Pranjal Protim Borah. 2020-07-28 11:55:47 +05:30
Prabhu Joseph
247eb0979b YARN-10319. Record Last N Scheduler Activities from ActivitiesManager
Reviewed by Tao Yang and Adam Antal.
2020-07-24 11:43:31 +05:30
bibinchundatt
bfcd775381 YARN-10315. Avoid sending RMNodeResourceupdate event if resource is same. Contributed by Sushil Ks. 2020-07-23 16:41:35 +05:30
Ayush Saxena
6bcb24d269 HADOOP-17100. Replace Guava Supplier with Java8+ Supplier in Hadoop. Contributed by Ahmed Hussein. 2020-07-18 14:33:43 +05:30
Bilwa S T
8980c23fc2
YARN-10350. TestUserGroupMappingPlacementRule fails
Signed-off-by: Akira Ajisaka <aajisaka@apache.org>
2020-07-15 14:31:34 +09:00
Eric Badger
48f90115b5 YARN-10348. Allow RM to always cancel tokens after app completes. Contributed by
Jim Brennan
2020-07-13 23:10:39 +00:00
Eric E Payne
0427100b75 YARN-10297. TestContinuousScheduling#testFairSchedulerContinuousSchedulingInitTime fails intermittently. Contributed by Jim Brennan (Jim_Brennan) 2020-07-13 18:57:50 +00:00
Xiaoyu Yao
f91a8ad88b
HADOOP-17079. Optimize UGI#getGroups by adding UGI#getGroupsSet. (#2085) 2020-07-09 11:33:37 -07:00
Prabhu Joseph
3a4d05b850 YARN-8047. RMWebApp make external class pluggable.
Contributed by Bilwa S T.
2020-07-08 12:43:20 +05:30
Prabhu Joseph
2bbd00dff4 YARN-10337. Fix failing testcase TestRMHATimelineCollectors.
Contributed by Bilwa S T.
2020-07-07 17:10:59 +05:30
Akira Ajisaka
639acb6d89
HADOOP-17111. Replace Guava Optional with Java8+ Optional. Contributed by Ahmed Hussein. 2020-07-06 16:08:36 +09:00
Szilard Nemeth
04abd0eb17 YARN-10330. Add missing test scenarios to TestUserGroupMappingPlacementRule and TestAppNameMappingPlacementRule. Contributed by Peter Bacsko 2020-07-01 14:10:55 +02:00
Eric Yang
e8dc862d38 YARN-9809. Added node manager health status to resource manager registration call.
Contributed by Eric Badger via eyang
2020-06-30 11:39:16 -07:00
Eric E Payne
74fc13cf91 YARN-9903: Support reservations continue looking for Node Labels. Contributed by Jim Brennan (Jim_Brennan). 2020-06-29 18:39:53 +00:00
Szilard Nemeth
33d3df7ee2 YARN-10277. CapacityScheduler test TestUserGroupMappingPlacementRule should build proper hierarchy. Contributed by Szilard Nemeth 2020-06-26 11:14:24 +02:00
Szilard Nemeth
6a8fd73b27 YARN-10279. Avoid unnecessary QueueMappingEntity creations. Contributed by Marton Hudaky 2020-06-25 17:27:58 +02:00
Szilard Nemeth
03f855e3e7 YARN-10316. FS-CS converter: convert maxAppsDefault, maxRunningApps settings. Contributed by Peter Bacsko 2020-06-23 12:13:04 +02:00
Prabhu Joseph
ce1008fe61 YARN-10321. Break down TestUserGroupMappingPlacementRule#testMapping into test scenarios
Contributed by Szilard Nemeth. Reviewed by Gergely Pollak.
2020-06-22 12:49:14 +05:30
Surendra Singh Lilhore
b2facc84a1 YARN-9460. QueueACLsManager and ReservationsACLManager should not use instanceof checks. Contributed by Bilwa S T. 2020-06-20 20:09:31 +05:30
Szilard Nemeth
469841446f YARN-9930. Support max running app logic for CapacityScheduler. Contributed by Peter Bacsko 2020-06-19 14:50:24 +02:00
Szilard Nemeth
5b1a56f9f1 YARN-10281. Redundant QueuePath usage in UserGroupMappingPlacementRule and AppNameMappingPlacementRule. Contributed by Gergely Pollak 2020-06-17 14:34:40 +02:00
Vinayakumar B
e154084770
HADOOP-17046. Support downstreams' existing Hadoop-rpc implementations using non-shaded protobuf classes (#2026) 2020-06-12 23:16:33 +05:30
Tao Yang
7c4de59fc1 YARN-10293. Reserved Containers not allocated from available space of other nodes in CandidateNodeSet for MultiNodePlacement. Contributed by Prabhu Joseph. 2020-06-12 22:10:25 +08:00