Commit Graph

2030 Commits

Author SHA1 Message Date
Wangda Tan
60f9e60b3b Preparing for 3.2.0 development
Change-Id: I6d0e01f3d665d26573ef2b957add1cf0cddf7938
2018-02-11 11:17:38 +08:00
Yufei Gu
1bc03ddf97 YARN-7655. Avoid AM preemption caused by RRs for specific nodes or racks. Contributed by Steven Rand. 2018-02-08 12:32:43 -08:00
Robert Kanter
814d701d46 YARN-7841. Cleanup AllocationFileLoaderService's reloadAllocations method (snemeth via rkanter) 2018-02-06 14:36:49 -08:00
Weiwei Yang
bff858e910 YARN-7860. Fix UT failure TestRMWebServiceAppsNodelabel#testAppsRunning. Contributed by Sunil G. 2018-02-06 19:33:04 +08:00
Konstantinos Karanasos
50723889cc YARN-7778. Merging of placement constraints defined at different levels. Contributed by Weiwei Yang. 2018-02-02 14:46:20 -08:00
Arun Suresh
6e5ba9366f YARN-7839. Modify PlacementAlgorithm to Check node capacity before placing request on node. (Panagiotis Garefalakis via asuresh) 2018-02-02 10:28:22 -08:00
Arun Suresh
d481344783 YARN-7822. Constraint satisfaction checker support for composite OR and AND constraints. (Weiwei Yang via asuresh) 2018-01-31 01:51:08 -08:00
Arun Suresh
add993e26a YARN-6597. Add RMContainer recovery test to verify tag population in the AllocationTagsManager. (Panagiotis Garefalakis via asuresh) 2018-01-31 01:30:17 -08:00
Arun Suresh
f8c5f5b237 YARN-7784. Fix Cluster metrics when placement processor is enabled. (asuresh) 2018-01-31 01:30:17 -08:00
Arun Suresh
c23980c4f2 YARN-7795. Fix jenkins issues of YARN-6592 branch. (Sunil G via asuresh) 2018-01-31 01:30:17 -08:00
Arun Suresh
644afe5fd8 YARN-7807. Assume intra-app anti-affinity as default for scheduling request inside AppPlacementAllocator. (Wangda Tan via asuresh) 2018-01-31 01:30:17 -08:00
Arun Suresh
a4c539fcdb YARN-7783. Add validation step to ensure constraints are not violated due to order in which a request is processed. (asuresh) 2018-01-31 01:30:17 -08:00
Sunil G
9b81cb0537 YARN-7779. Display allocation tags in RM web UI and expose same through REST API. Contributed by Weiwei Yang. 2018-01-31 01:30:17 -08:00
Konstantinos Karanasos
adbe87abf8 YARN-7788. Factor out management of temp tags from AllocationTagsManager. (Arun Suresh via kkaranasos) 2018-01-31 01:30:17 -08:00
Arun Suresh
8bf7c44436 YARN-7763. Allow Constraints specified in the SchedulingRequest to override application level constraints. (Weiwei Yang via asuresh) 2018-01-31 01:30:17 -08:00
Arun Suresh
28fe7f3318 YARN-7774. Miscellaneous fixes to the PlacementProcessor. (asuresh) 2018-01-31 01:30:17 -08:00
Arun Suresh
38af237969 YARN-6599. Support anti-affinity constraint via AppPlacementAllocator. (Wangda Tan via asuresh) 2018-01-31 01:30:17 -08:00
Arun Suresh
8779a35742 YARN-7709. Remove SELF from TargetExpression type. (Konstantinos Karanasos via asuresh) 2018-01-31 01:30:17 -08:00
Wangda Tan
29d9e4d581 YARN-6619. AMRMClient Changes to use the PlacementConstraint and SchcedulingRequest objects. (Arun Suresh via wangda) 2018-01-31 01:30:17 -08:00
Arun Suresh
a5c1fc881e YARN-7696. Add container tags to ContainerTokenIdentifier, api.Container and NMContainerStatus to handle all recovery cases. (asuresh) 2018-01-31 01:30:17 -08:00
Arun Suresh
4eda58c136 YARN-7681. Double-check placement constraints in scheduling phase before actual allocation is made. (Weiwei Yang via asuresh) 2018-01-31 01:30:17 -08:00
Arun Suresh
bdba01f73b YARN-7682. Expose canSatisfyConstraints utility function to validate a placement against a constraint. (Panagiotis Garefalakis via asuresh) 2018-01-31 01:30:17 -08:00
Arun Suresh
a52d11fb8c YARN-7613. Implement Basic algorithm for constraint based placement. (Panagiotis Garefalakis via asuresh) 2018-01-31 01:30:17 -08:00
Arun Suresh
f9af15d659 YARN-7612. Add Processor Framework for Rich Placement Constraints. (asuresh) 2018-01-31 01:30:17 -08:00
Arun Suresh
1efb2b6f25 YARN-6596. Introduce Placement Constraint Manager module. (Konstantinos Karanasos via asuresh) 2018-01-31 01:30:17 -08:00
Arun Suresh
37f1a7b64f YARN-7653. Node group support for AllocationTagsManager. (Panagiotis Garefalakis via asuresh) 2018-01-31 01:30:17 -08:00
Arun Suresh
06eb63e64b YARN-7669. API and interface modifications for placement constraint processor. (asuresh) 2018-01-31 01:30:17 -08:00
Arun Suresh
88d8d3f40b YARN-7670. Modifications to the ResourceScheduler API to support SchedulingRequests. (asuresh) 2018-01-31 01:30:17 -08:00
Arun Suresh
801c0988b5 YARN-7522. Introduce AllocationTagsManager to associate allocation tags to nodes. (Wangda Tan via asuresh) 2018-01-31 01:30:17 -08:00
Sunil G
e9c72d04be YARN-7790. Improve Capacity Scheduler Async Scheduling to better handle node failures. Contributed by Wangda Tan. 2018-01-29 20:44:38 +05:30
Jinjiang Ling
3400d0c535
YARN-7698. A misleading variable's name in ApplicationAttemptEventDispatcher
Signed-off-by: Akira Ajisaka <aajisaka@apache.org>
2018-01-30 00:00:57 +09:00
Wangda Tan
75618c0dec YARN-7760. [UI2] Clicking 'Master Node' or link next to 'AM Node Web UI' under application's appAttempt page goes to OLD RM UI. (Vasudevan Skm / Sunil G via wangda)
Change-Id: Ia7e2b955d3551ca43c6852fe253fd01983b56565
2018-01-26 15:44:55 +08:00
Wangda Tan
e0cfb0a31a YARN-7817. Add Resource reference to RM's NodeInfo object so REST API can get non memory/vcore resource usages. (Sunil G via wangda)
Change-Id: Ia7ceeabd82046645ddeaf487c763288f36cfbdee
2018-01-26 15:43:27 +08:00
Jason Lowe
ff8378eb1b YARN-7102. NM heartbeat stuck when responseId overflows MAX_INT. Contributed by Botong Huang 2018-01-25 17:47:19 -06:00
Sunil G
82cc6f6968 YARN-7728. Expose container preemptions related information in Capacity Scheduler queue metrics. Contributed by Eric Payne. 2018-01-25 14:32:54 +05:30
Billie Rinaldi
e307edcb47 YARN-7540 and YARN-7605. Convert yarn app cli to call yarn api services and implement doAs for Api Service REST API. Contributed by Eric Yang and Jian He 2018-01-23 17:54:39 -08:00
Yufei Gu
22ee6f77e5 YARN-7755. Clean up deprecation messages for allocation increments in FS config. Contributed by Wilfred Spiegelenburg. 2018-01-22 10:30:01 -08:00
Miklos Szegedi
bc93ac229e YARN-7139. FairScheduler: finished applications are always restored to default queue. Contributed by Wilfred Spiegelenburg. 2018-01-18 17:43:47 -08:00
Sunil G
06cceba1cb YARN-7738. CapacityScheduler: Support refresh maximum allocation for multiple resource types. Contributed by Wangda Tan. 2018-01-18 19:05:26 +05:30
Yufei Gu
370f1c6283 YARN-6486. FairScheduler: Deprecate continuous scheduling. (Contributed by Wilfred Spiegelenburg) 2018-01-16 14:46:55 -08:00
Rohith Sharma K S
d09058b2fd YARN-6736. Consider writing to both ats v1 & v2 from RM for smoother upgrades. Contributed by Aaron Gresch. 2018-01-16 07:58:29 +05:30
Sunil G
128d773a23 YARN-7727. Incorrect log levels in few logs with QueuePriorityContainerCandidateSelector. Contributed by Prabhu Joseph. 2018-01-12 15:41:19 +05:30
Sunil G
afd8caba27 YARN-7722. Rename variables in MockNM, MockRM for better clarity. Contributed by Lovekesh bansal 2018-01-10 20:51:34 +05:30
Miklos Szegedi
55066cc53d YARN-7689. TestRMContainerAllocator fails after YARN-6124. Contributed by Wilfred Spiegelenburg. 2018-01-09 16:59:39 -08:00
Haibo Chen
8ee7080e5d YARN-7665. Allow FS scheduler state dump to be turned on/off separately from FS debug log. (Wilfred Spiegelenburg via Haibo Chen) 2018-01-09 16:45:35 -08:00
Jason Lowe
90c86d6e09 YARN-6948. Invalid event: ATTEMPT_ADDED at FINAL_SAVING. Contributed by lujie 2018-01-09 10:51:15 -06:00
Jason Lowe
ebff4decfe YARN-7663. RMAppImpl:Invalid event: START at KILLED. Contributed by lujie 2018-01-09 09:52:59 -06:00
bibinchundatt
b26e30ab16 YARN-7508. NPE in FiCaSchedulerApp when debug log enabled in async-scheduling mode. Contributed by Tao Yang. 2018-01-09 17:57:01 +05:30
Ray Chiang
59ab5da0a0 YARN-4227. Ignore expired containers from removed nodes in FairScheduler. (Wilfred Spiegelenburg via rchiang) 2018-01-08 16:19:02 -08:00
Rohith Sharma K S
c2d6fa3656 YARN-7699. queueUsagePercentage is coming as INF for getApp REST api call. Contributed by Sunil G. 2018-01-08 10:39:31 +05:30
Wangda Tan
a81144daa0 YARN-7666. Introduce scheduler specific environment variable support in ApplicationSubmissionContext for better scheduling placement configurations. (Sunil G via wangda)
Change-Id: I0fd826490f5160d47d42af2a9ac0bd8ec4e959dc
2018-01-05 15:12:04 -08:00
Robert Kanter
2aa4f0a559 YARN-7645. TestContainerResourceUsage#testUsageAfterAMRestartWithMultipleContainers is flakey with FairScheduler (rkanter) 2018-01-05 13:55:09 -08:00
Robert Kanter
f8e7dd9b10 YARN-7557. It should be possible to specify resource types in the fair scheduler increment value (grepas via rkanter) 2018-01-05 11:15:06 -08:00
Sunil G
0c75d0634b YARN-7619. Max AM Resource value in Capacity Scheduler UI has to be refreshed for every user. Contributed by Eric Payne. 2018-01-05 14:42:17 +05:30
Robert Kanter
7a55044803 YARN-7622. Allow fair-scheduler configuration on HDFS (gphillips via rkanter) 2018-01-03 15:31:50 -08:00
Rohith Sharma K S
c9bf813c9a YARN-7692. Skip validating priority acls while recovering applications. Contributed by Sunil G. 2018-01-03 18:20:04 +05:30
Robert Kanter
382215c72b YARN-7577. Unit Fail: TestAMRestart#testPreemptedAMRestartOnRMRestart (miklos.szegedi@cloudera.com via rkanter) 2017-12-20 13:39:00 -08:00
Eric Yang
94a2ac6b71 YARN-7466. addendum patch for failing unit test. (Contributed by Chandni Singh) 2017-12-19 18:42:27 -05:00
Akira Ajisaka
001008958d
YARN-7664. Several javadoc errors. Contributed by Sean Mackrory. 2017-12-18 22:24:51 +09:00
Rohith Sharma K S
09d996fdd4 YARN-7190. Ensure only NM classpath in 2.x gets TSv2 related hbase jars, not the user classpath. Contributed by Varun Saxena. 2017-12-15 21:50:28 +05:30
Sunil G
890d3d0645 YARN-7638. Unit tests related to preemption for auto created leaf queues feature.Contributed by Suma Shivaprasad. 2017-12-15 13:00:57 +05:30
Sunil G
cb87e4dc92 YARN-7643. Handle recovery of applications in case of auto-created leaf queue mapping. Contributed by Suma Shivaprasad. 2017-12-13 22:49:58 +05:30
Sunil G
8bb83a8f62 Queue ACL validations should validate parent queue ACLs before auto-creating leaf queues. Contributed by Suma Shivaprasad. 2017-12-12 15:20:59 +05:30
Sunil G
5c87fb2f62 YARN-7635. TestRMWebServicesSchedulerActivities fails in trunk. Contributed by Sunil G. 2017-12-12 15:08:18 +05:30
Sunil G
312ceebde8 YARN-7632. Effective min and max resource need to be set for auto created leaf queues upon creation and capacity management. Contributed by Suma Shivaprasad. 2017-12-11 19:20:02 +05:30
Wangda Tan
adca1a72e4 YARN-7591. NPE in async-scheduling mode of CapacityScheduler. (Tao Yang via wangda)
Change-Id: I46689e530550ee0a6ac7a29786aab2cc1bdf314f
2017-12-08 15:17:02 -08:00
Wangda Tan
a8316df8c0 YARN-7520. Queue Ordering policy changes for ordering auto created leaf queues within Managed parent Queues. (Suma Shivaprasad via wangda)
Change-Id: I482f086945bd448d512cb5b3879d7371e37ee134
2017-12-08 15:11:28 -08:00
Wangda Tan
f548bfffbd YARN-7420. YARN UI changes to depict auto created queues. (Suma Shivaprasad via wangda)
Change-Id: I8039d3772a191ddede132cd1f8b08a8ca2e275b7
2017-12-08 15:10:47 -08:00
Wangda Tan
b38643c9a8 YARN-7473. Implement Framework and policy for capacity management of auto created queues. (Suma Shivaprasad via wangda)
Change-Id: Icca7805fe12f6f7fb335effff4b121b6f7f6337b
2017-12-08 15:10:16 -08:00
Wangda Tan
74665e3a7d YARN-7274. Ability to disable elasticity at leaf queue level. (Zian Chen via wangda)
Change-Id: Ic8d43e297f0f5de788b562f7eff8106c5c35e8d2
2017-12-08 15:07:56 -08:00
Sunil G
4db4a4a165 YARN-7575. NPE in scheduler UI when max-capacity is not configured. Contributed by Sunil G. 2017-12-07 18:56:54 -08:00
Sunil G
daa1cdd062 YARN-7564. Cleanup to fix checkstyle issues of YARN-5881 branch. Contributed by Sunil G. 2017-12-07 18:56:54 -08:00
Wangda Tan
1012b901c8 YARN-7544. Use queue-path.capacity/maximum-capacity to specify absolute min/max resources. (Sunil G via wangda)
Change-Id: I685341be213eee500f51e02f01c91def89391c17
2017-12-07 18:56:54 -08:00
Wangda Tan
b7b8cd5324 YARN-7538. Fix performance regression introduced by Capacity Scheduler absolute min/max resource refactoring. (Sunil G via wangda)
Change-Id: Ic9bd7e599c56970fe01cb0e1bba6df7d1f77eb29
2017-12-07 18:56:54 -08:00
Wangda Tan
7462c38277 YARN-7483. CapacityScheduler test cases cleanup post YARN-5881. (Sunil G via wangda)
Change-Id: I9741a6baf5cb7352d05636efb6c0b24790e7589a
2017-12-07 18:56:54 -08:00
Rohith Sharma K S
e65ca92fb6 YARN-7482. Max applications calculation per queue has to be retrospected with absolute resource support. Contributed by Sunil G. 2017-12-07 18:56:54 -08:00
Wangda Tan
034b312d9f YARN-7411. Inter-Queue preemption's computeFixpointAllocation need to handle absolute resources while computing normalizedGuarantee. (Sunil G via wangda)
Change-Id: I41b1d7558c20fc4eb2050d40134175a2ef6330cb
2017-12-07 18:56:54 -08:00
Wangda Tan
aa3f62740f YARN-7332. Compute effectiveCapacity per each resource vector. (Sunil G via wangda) 2017-12-07 18:56:54 -08:00
Wangda Tan
d52627a7cb YARN-7254. UI and metrics changes related to absolute resource configuration. (Sunil G via wangda) 2017-12-07 18:56:54 -08:00
Wangda Tan
5e798b1a0d YARN-6471. Support to add min/max resource configuration for a queue. (Sunil G via wangda)
Change-Id: I9213f5297a6841fab5c573e85ee4c4e5f4a0b7ff
2017-12-07 18:56:54 -08:00
Sunil G
a957f1c60e YARN-7438. Additional changes to make SchedulingPlacementSet agnostic to ResourceRequest / placement algorithm. Contributed by Wangda Tan 2017-12-05 22:50:07 +05:30
Sunil G
f9f317b702 YARN-7586. Application Placement should be done before ACL checks in ResourceManager. Contributed by Suma Shivaprasad. 2017-12-05 18:28:31 +05:30
Robert Kanter
d8863fc16f YARN-5594. Handle old RMDelegationToken format when recovering RM (rkanter) 2017-12-04 13:14:55 -08:00
Arun Suresh
37ca416950 YARN-7587. Skip dispatching opportunistic containers to nodes whose queue is already full. (Weiwei Yang via asuresh) 2017-12-03 22:22:01 -08:00
Robert Kanter
c83fe44917 YARN-4813. TestRMWebServicesDelegationTokenAuthentication.testDoAs fails intermittently (grepas via rkanter) 2017-12-01 12:18:13 -08:00
Wangda Tan
a63d19d365 YARN-6124. Make SchedulingEditPolicy can be enabled / disabled / updated with RMAdmin -refreshQueues. (Zian Chen via wangda)
Change-Id: Id93656f3af7dcd78cafa94e33663c78d410d43c2
2017-11-30 15:57:22 -08:00
Daniel Templeton
8498d287cd YARN-7541. Node updates don't update the maximum cluster capability for resources other than CPU and memory 2017-11-29 11:11:36 -08:00
Jason Lowe
a2c7a73e33 YARN-6647. RM can crash during transitionToStandby due to InterruptedException. Contributed by Bibin A Chundatt 2017-11-28 11:15:44 -06:00
Jian He
fedabcad42 YARN-6168. Restarted RM may not inform AM about all existing containers. Contributed by Chandni Singh 2017-11-27 10:19:58 -08:00
Yufei Gu
2bde3aedf1 YARN-7290. Method canContainerBePreempted can return true when it shouldn't. (Contributed by Steven Rand) 2017-11-24 23:32:46 -08:00
Wangda Tan
834e91ee91 YARN-7509. AsyncScheduleThread and ResourceCommitterService are still running after RM is transitioned to standby. (Tao Yang via wangda)
Change-Id: I7477fe355419fd4a0a6e2bdda7319abad4c4c748
2017-11-23 19:59:03 -08:00
Arun Suresh
b46ca7e73b YARN-6483. Add nodes transitioning to DECOMMISSIONING state to the list of updated nodes returned to the AM. (Juan Rodriguez Hortala via asuresh) 2017-11-22 19:18:30 -08:00
Yufei Gu
4cc9479dae YARN-7524. Remove unused FairSchedulerEventLog. (Contributed by Wilfred Spiegelenburg) 2017-11-22 14:18:36 -08:00
yufei
03c311eae3 YARN-7513. Remove the scheduler lock in FSAppAttempt.getWeight() (Contributed by Wilfred Spiegelenburg) 2017-11-21 10:33:34 -08:00
Wangda Tan
0d781dd03b YARN-7527. Over-allocate node resource in async-scheduling mode of CapacityScheduler. (Tao Yang via wangda)
Change-Id: I51ae6c2ab7a3d1febdd7d8d0519b63a13295ac7d
2017-11-20 11:48:15 -08:00
bibinchundatt
b5b81a4f08 YARN-7489. ConcurrentModificationException in RMAppImpl#getRMAppMetrics. Contributed by Tao Yang. 2017-11-18 19:25:29 +05:30
Eric Yang
0940e4f692 YARN-7218. Decouple YARN Services REST API namespace from RM. (Contributed by Eric Yang) 2017-11-17 12:28:12 -05:00
Wangda Tan
0987a7b8cb YARN-7419. CapacityScheduler: Allow auto leaf queue creation after queue mapping. (Suma Shivaprasad via wangda)
Change-Id: Ia1704bb8cb5070e5b180b5a85787d7b9ca57ebc6
2017-11-16 11:25:52 -08:00
Sunil G
61ace174cd YARN-7469. Capacity Scheduler Intra-queue preemption: User can starve if newest app is exactly at user limit. Contributed by Eric Payne. 2017-11-16 22:34:23 +05:30
Daniel Templeton
b246c54749 YARN-7414. FairScheduler#getAppWeight() should be moved into FSAppAttempt#getWeight()
(Contributed by Soumabrata Chakraborty via Daniel Templeton)
2017-11-15 10:03:29 -08:00