big-data/hadoop
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
16,179 Commits 2 Branches 2 Tags 264 MiB
0615985886
Commit Graph

1240 Commits

Author SHA1 Message Date
Jason Lowe
d7762a5511 YARN-6165. Intra-queue preemption occurs even when preemption is turned off for a specific queue. Contributed by Eric Payne 2017-03-08 16:46:09 -06:00
Sunil G
1eb8186703 YARN-6207. Move application across queues should handle delayed event processing. Contributed by Bibin A Chundatt. 2017-03-08 12:04:30 +05:30
Jian He
235203dffd YARN-6153. KeepContainer does not work when AM retry window is set. Contributed by kyungwan nam 2017-02-28 13:23:36 -08:00
Xuan
e0bb867c3f YARN-6189: Improve application status log message when RM restarted when 
app is in NEW state. Contributed by Junping Du
 2017-02-28 11:04:56 -08:00
Wangda Tan
eac6b4c35c YARN-6216. Unify Container Resizing code paths with Container Updates making it scheduler agnostic. (Arun Suresh via wangda) 2017-02-28 10:35:50 -08:00
Karthik Kambatla
f187d63816 YARN-6231. FairSchedulerTestBase helper methods should call scheduler.update to avoid flakiness. (kasha) 2017-02-27 22:10:57 -08:00
Naganarasimha
5f5b031d1f YARN-5703. ReservationAgents are not correctly configured. Contributed by Manikandan R. 2017-02-27 20:38:29 +05:30
Karthik Kambatla
fbfe86deea YARN-6172. FSLeafQueue demand update needs to be atomic. (Miklos Szegedi via kasha) 2017-02-26 20:36:33 -08:00
Karthik Kambatla
694e680d20 YARN-6222. TestFairScheduler.testReservationMetrics is flaky. (Yufei Gu via kasha) 2017-02-23 15:21:52 -08:00
Karthik Kambatla
b10e962224 YARN-6194. Cluster capacity in SchedulingPolicy is updated only on allocation file reload. (Yufei Gu via kasha) 2017-02-22 15:58:49 -08:00
Karthik Kambatla
718ad9f6ee YARN-6210. FairScheduler: Node reservations can interfere with preemption. (kasha) 2017-02-22 15:46:07 -08:00
Wangda Tan
1a6ca75f38 YARN-6143. Fix incompatible issue caused by YARN-3583. (Sunil G via wangda) 2017-02-22 11:17:09 -08:00
Karthik Kambatla
dbbfcf74ab YARN-6193. FairScheduler might not trigger preemption when using DRF. (kasha) 2017-02-17 14:07:31 -08:00
Daniel Templeton
c7a36e6130 YARN-6125. The application attempt's diagnostic message should have a maximum size 
(Contributed by Andras Piros via Daniel Templeton)
 2017-02-17 13:40:58 -08:00
Rohith Sharma K S
a393e84c6f YARN-6200. Reverting since the same functionality achieved by YARN-1623. 
Revert "YARN-5068. Expose scheduler queue to application master. (Harish Jaiprakash via rohithsharmaks)"

This reverts commit b7ac85259c.
 2017-02-16 18:00:25 +05:30
Karthik Kambatla
11be3f70e0 YARN-4212. FairScheduler: Can't create a DRF queue under a FAIR policy queue. (Yufei Gu via kasha) 2017-02-15 23:51:22 -08:00
Karthik Kambatla
6c25dbcdc0 YARN-6163. FS Preemption is a trickle for severely starved applications. (kasha) 2017-02-15 23:16:12 -08:00
Varun Saxena
b7613e0f40 YARN-6156. AM blacklisting to consider node label partition (Bibin A Chundatt via Varun Saxena) 2017-02-15 14:48:17 +05:30
Arun Suresh
aaf106fde3 YARN-5966. AMRMClient changes to support ExecutionType update. (asuresh) 2017-02-14 06:09:10 -08:00
Karthik Kambatla
646c6d6509 YARN-3933. FairScheduler: Multiple calls to completedContainer are not safe. (Shiwei Guo and Miklos Szegedi via kasha) 2017-02-13 11:26:30 -08:00
Karthik Kambatla
08f93978f3 YARN-6144. FairScheduler: preempted resources can become negative. (Miklos Szegedi via kasha) 2017-02-09 14:45:04 -08:00
Wangda Tan
5fb723bb77 YARN-5889. Improve and refactor user-limit calculation in Capacity Scheduler. (Sunil G via wangda) 2017-02-09 10:23:50 -08:00
Daniel Templeton
3e06475307 YARN-4658. Typo in o.a.h.yarn.server.resourcemanager.scheduler.fair.TestFairScheduler comment (Contributed by Udai Potluri via Daniel Templeton) 2017-01-31 16:25:33 -08:00
Karthik Kambatla
11e44bdda7 YARN-6131. FairScheduler: Lower update interval for faster tests. (kasha) 2017-01-30 13:35:23 -08:00
Sunil G
165f07f51a YARN-6123. [YARN-5864] Add a test to make sure queues of orderingPolicy will be updated when childQueues is added or removed. Contributed by Wangda Tan. 2017-01-27 18:37:51 +05:30
Daniel Templeton
f85b74ccf9 YARN-4975. Fair Scheduler: exception thrown when a parent queue marked 'parent' has configured child queues 
(Contributed by Yufei Gu via Daniel Templeton)
 2017-01-26 10:31:09 -08:00
Karthik Kambatla
abedb8a9d8 YARN-5830. FairScheduler: Avoid preempting AM containers. (Yufei Gu via kasha) 2017-01-25 12:17:28 -08:00
Rohith Sharma K S
0101267d9d YARN-6082. Invalid REST api response for getApps since queueUsagePercentage is coming as INF. Contributed by Sunil G. 2017-01-24 10:59:55 +05:30
Wangda Tan
ce832059db YARN-5864. Capacity Scheduler - Queue Priorities. (wangda) 2017-01-23 10:52:14 -08:00
Jason Lowe
69fa81679f YARN-5910. Support for multi-cluster delegation tokens. Contributed by Jian He 2017-01-23 17:12:51 +00:00
Sunil G
3fa0d540df YARN-6031. Application recovery has failed when node label feature is turned off during RM recovery. Contributed by Ying Zhang. 2017-01-23 12:19:35 +05:30
Sangjin Lee
b01514f65b YARN-5928. Move ATSv2 HBase backend code into a new module that is only dependent at runtime by yarn servers. Contributed by Haibo Chen. 2017-01-19 20:52:55 -08:00
Wangda Tan
72054a817d YARN-5556. CapacityScheduler: Support deleting queues without requiring a RM restart. (Naganarasimha G R via wangda) 2017-01-18 13:31:33 -08:00
Karthik Kambatla
e224c96234 YARN-5831. FairScheduler: Propagate allowPreemptionFrom flag all the way down to the app. (Yufei Gu via kasha) 2017-01-17 17:01:31 -08:00
Daniel Templeton
695f110f93 YARN-6071. Fix incompatible API change on AM-RM protocol due to YARN-3866 (trunk only) 
(Contributed by Wangda Tan via Daniel Templeton)
 2017-01-17 12:55:47 -08:00
Sunil G
d3170f9eba YARN-6081. LeafQueue#getTotalPendingResourcesConsideringUserLimit should deduct reserved from pending to avoid unnecessary preemption of reserved container. Contributed by Wangda Tan. 2017-01-13 18:22:29 +05:30
Daniel Templeton
7979939428 YARN-5554. MoveApplicationAcrossQueues does not check user permission on the target queue 
(Contributed by Wilfred Spiegelenburg via Daniel Templeton)
 2017-01-11 14:34:10 -08:00
Jason Lowe
357eab9566 YARN-5416. TestRMRestart#testRMRestartWaitForPreviousAMToFinish failed intermittently due to not wait SchedulerApplicationAttempt to be stopped. Contributed by Junping Du 2017-01-11 17:13:37 +00:00
Junping Du
4db119b7b5 YARN-6079. Fix simple spelling errors in yarn test code. Contributed by vijay. 2017-01-10 18:20:27 -08:00
Junping Du
945db55f2e YARN-4148. When killing app, RM releases app's resource before they are released by NM. Contributed by Jason Lowe. 2017-01-09 18:14:46 -08:00
Wangda Tan
287d3d6804 YARN-3955. Support for application priority ACLs in queues of CapacityScheduler. (Sunil G via wangda) 2017-01-09 08:40:39 -08:00
Arun Suresh
2977bc6a14 YARN-6040. Introduce api independent PendingAsk to replace usage of ResourceRequest within Scheduler classes. (Wangda Tan via asuresh) 2017-01-06 09:59:57 -08:00
Wangda Tan
0a55bd841e YARN-5959. RM changes to support change of container ExecutionType. (Arun Suresh via wangda) 2017-01-05 10:31:05 -08:00
Rohith Sharma K S
e49e0a6e37 YARN-5988. RM unable to start in secure setup. Contributed by Ajith S. 2017-01-04 12:09:29 +05:30
Sunil G
9ca54f4810 YARN-5906. Update AppSchedulingInfo to use SchedulingPlacementSet. Contributed by Wangda Tan. 2016-12-28 22:48:19 +05:30
Sunil G
f6715b26b6 YARN-6024. Capacity Scheduler 'continuous reservation looking' doesn't work when sum of queue's used and reserved resources is equal to max. Contributed by Wangda Tan. 2016-12-28 12:33:02 +05:30
Wangda Tan
0840b4329b YARN-5756. Add state-machine implementation for scheduler queues. (Xuan Gong via wangda) 2016-12-27 21:18:33 -08:00
Robert Kanter
1bbd023275 YARN-5962. Spelling errors in logging and exceptions for resource manager code (gsohn via rkanter) 2016-12-27 13:35:34 -08:00
Arun Suresh
ac1e5d4f77 YARN-5938. Refactoring OpportunisticContainerAllocator to use SchedulerRequestKey instead of Priority and other misc fixes (asuresh) 2016-12-27 12:40:44 -08:00
Arun Suresh
2273a74c1f YARN-5646. Add documentation and update config parameter names for scheduling of OPPORTUNISTIC containers. (Konstantinos Karanasos via asuresh) 2016-12-16 08:14:34 -08:00
Robert Kanter
4b149a1e77 YARN-4457. Cleanup unchecked types for EventHandler (templedf via rkanter) 2016-12-09 17:28:25 -08:00
Jian He
a6410a542e YARN-5709. Cleanup leader election configs and pluggability. Contribtued by Karthik Kambatla 2016-12-09 17:00:37 -08:00
Arun Suresh
b0aace21b1 YARN-5982. Simplify opportunistic container parameters and metrics. (Konstantinos Karanasos via asuresh) 2016-12-09 16:41:25 -08:00
Sunil G
74d0066d33 YARN-5970. Validate application update timeout request parameters. Contributed by Rohith Sharma K S. 2016-12-08 15:53:56 +05:30
Daniel Templeton
9f5d2c4fff YARN-5136. Error in handling event type APP_ATTEMPT_REMOVED to the scheduler 
(Contributed by Wilfred Spiegelenburg via Daniel Templeton)
 2016-12-07 11:12:14 -08:00
Sunil
ab923a53fc YARN-5965. Retrospect ApplicationReport#getApplicationTimeouts. Contributed by Rohith Sharma K S 2016-12-08 00:27:25 +05:30
Rohith Sharma K S
563480dccd YARN-5932. Retrospect moveApplicationToQueue in align with YARN-5611. Contributed by Sunil G. 2016-12-07 10:39:14 +05:30
Jian He
43ebff2e35 YARN-5559. Analyse 2.8.0/3.0.0 jdiff reports and fix any issues. Contributed by Akira Ajisaka & Wangda Tan 2016-12-05 11:39:34 -08:00
Jian He
f885160f4a YARN-5746. The state of the parentQueue and its childQueues should be synchronized. Contributed by Xuan Gong 2016-12-02 16:17:31 -08:00
Daniel Templeton
5bd18c49bd YARN-5929. Missing scheduling policy in the FS queue metric. (Contributed by Yufei Gu via Daniel Templeton) 2016-12-02 13:55:42 -08:00
Li Lu
69fb70c31a YARN-5761. Separate QueueManager from Scheduler. (Xuan Gong via gtcarrera9) 2016-11-30 13:38:42 -08:00
Daniel Templeton
b3befc021b YARN-4997. Update fair scheduler to use pluggable auth provider (Contributed by Tao Jie via Daniel Templeton) 2016-11-30 09:50:33 -08:00
Kai Zheng
cfd8076f81 Revert due to an error "HDFS-10994. Support an XOR policy XOR-2-1-64k in HDFS. Contributed by Sammi Chen" 
This reverts commit 5614f847b2ef2a5b70bd9a06edc4eba06174c6.
 2016-11-30 15:44:52 +08:00
Daniel Templeton
3b9d3acd20 YARN-5890. FairScheduler should log information about AM-resource-usage and max-AM-share for queues 
(Contributed by Yufei Gu via Daniel Templeton)
 2016-11-29 12:46:05 -08:00
Daniel Templeton
25f9872be6 YARN-5774. MR Job stuck in ACCEPTED status without any progress in Fair Scheduler 
if set yarn.scheduler.minimum-allocation-mb to 0. (Contributed by Yufei Gu via Daniel Templeton)
 2016-11-29 09:40:49 -08:00
Kai Zheng
5d5614f847 HDFS-10994. Support an XOR policy XOR-2-1-64k in HDFS. Contributed by Sammi Chen 2016-11-28 14:34:44 +08:00
Sunil
eb0a483ed0 YARN-4206. Add Application timeouts in Application report and CLI. Contributed by Rohith Sharma K S. 2016-11-24 18:18:42 +05:30
Rohith Sharma K S
e15c20edba YARN-5920. Fix deadlock in TestRMHA.testTransitionedToStandbyShouldNotHang. Contributed by Varun Saxena. 2016-11-24 12:18:38 +05:30
Daniel Templeton
10468529a9 YARN-4752. Improved preemption in FairScheduler. (kasha) 
Contains:
YARN-5605. Preempt containers (all on one node) to meet the requirement of starved applications
YARN-5821. Drop left-over preemption-related code and clean up method visibilities in the Schedulable hierarchy
YARN-5783. Verify identification of starved applications.
YARN-5819. Verify fairshare and minshare preemption
YARN-5885. Cleanup YARN-4752 branch for merge

Change-Id: Iee0962377d019dd64dc69a020725d2eaf360858c
 2016-11-23 19:48:59 -10:00
Jian He
1f12867a69 YARN-5649. Add REST endpoints for updating application timeouts. Contributed by Rohith Sharma K S 2016-11-23 16:25:39 -08:00
Arun Suresh
005850b28f YARN-5918. Handle Opportunistic scheduling allocate request failure when NM is lost. (Bibin A Chundatt via asuresh) 2016-11-23 09:53:31 -08:00
Rohith Sharma K S
a926f895c1 YARN-5865. Retrospect updateApplicationPriority api to handle state store exception in align with YARN-5611. Contributed by Sunil G. 2016-11-22 14:49:15 +05:30
Rohith Sharma K S
d65603517e YARN-5375. invoke MockRM#drainEvents implicitly in MockRM methods to reduce test failures. Contributed by sandflee. 2016-11-16 15:14:00 +05:30
Xiao Chen
f121d0b036 YARN-5875. TestTokenClientRMService#testTokenRenewalWrongUser fails. Contributed by Gergely Novák. 2016-11-15 13:58:11 -08:00
Rohith Sharma K S
b7070f3308 YARN-5874. RM -format-state-store and -remove-application-from-state-store commands fail with NPE. Contributed by Varun Saxena. 2016-11-15 10:58:25 +05:30
Jian He
fad9609d13 YARN-5825. ProportionalPreemptionalPolicy should use readLock over LeafQueue instead of synchronized block. Contributed by Sunil G 2016-11-11 15:16:21 -08:00
Naganarasimha
503e73e849 YARN-5545. Fix issues related to Max App in capacity scheduler. Contributed by Bibin A Chundatt 2016-11-11 20:48:31 +05:30
Eric Payne
93eeb13164 YARN-4218. Metric for resource*time that was preempted. Contributed by Chang Li. 2016-11-10 22:35:12 +00:00
Karthik Kambatla
86ac1ad9fd YARN-5453. FairScheduler#update may skip update demand resource of child queue/app if current demand reached maxResource. (sandflee via kasha) 2016-11-09 23:44:02 -08:00
Jian He
bcc15c6290 YARN-5611. Provide an API to update lifetime of an application. Contributed by Rohith Sharma K S 2016-11-09 16:08:05 -08:00
Naganarasimha
edbee9e609 YARN-4498. Application level node labels stats to be available in REST (addendum patch). Contributed by Bibin A Chundatt. 2016-11-10 05:00:05 +05:30
Daniel Templeton
59ee8b7a88 YARN-4329. [YARN-5437] Allow fetching exact reason as to why a submitted app 
is in ACCEPTED state in Fair Scheduler (Contributed by Yufei Gu)
 2016-11-09 13:11:37 -08:00
Jason Lowe
3f93ac0733 YARN-5356. NodeManager should communicate physical resource capability to ResourceManager. Contributed by Inigo Goiri 2016-11-08 22:01:26 +00:00
Jian He
de3b4aac56 YARN-5716. Add global scheduler interface definition and update CapacityScheduler to use it. Contributed by Wangda Tan 2016-11-07 10:14:39 -08:00
Jason Lowe
6bb741ff0e YARN-5837. NPE when getting node status of a decommissioned node after an RM restart. Contributed by Robert Kanter 2016-11-04 22:20:21 +00:00
Arun Suresh
0aafc122d4 YARN-2995. Enhance UI to show cluster resource utilization of various container Execution types. (Konstantinos Karanasos via asuresh) 2016-11-04 07:31:54 -07:00
Sunil
19b3779ae7 YARN-5802. updateApplicationPriority api in scheduler should ensure to re-insert app to correct ordering policy. Contributred by Bibin A Chundatt 2016-11-04 16:07:28 +05:30
Jason Lowe
352cbaa7a5 YARN-4862. Handle duplicate completed containers in RMNodeImpl. Contributed by Rohith Sharma K S 2016-11-03 13:54:31 +00:00
Varun Saxena
377919010b YARN-5815. Random failure of TestApplicationPriority.testOrderOfActivatingThePriorityApplicationOnRMRestart (Bibin A Chundatt via Varun Saxena) 2016-11-03 00:37:09 +05:30
Varun Saxena
7d2d8d25ba YARN-5788. Apps not activiated and AM limit resource in UI and REST not updated after -replaceLabelsOnNode (Bibin A Chundatt via Varun Saxena) 2016-11-01 15:32:04 +05:30
Wangda Tan
90dd3a8148 YARN-2009. CapacityScheduler: Add intra-queue preemption for app priority support. (Sunil G via wangda) 2016-10-31 15:18:31 -07:00
Daniel Templeton
cc2c993a8a YARN-4907. Make all MockRM#waitForState consistent. (Contributed by Yufei Gu via Daniel Templeton) 2016-10-31 13:20:56 -07:00
Naganarasimha
e0bebbbcdd YARN-4498. Application level node labels stats to be available in REST. Contributed by Bibin A Chundatt 2016-10-31 04:38:20 +05:30
Arun Suresh
aa3cab1eb2 YARN-5799. Fix Opportunistic Allocation to set the correct value of Node Http Address. (asuresh) 2016-10-29 02:03:57 -07:00
Varun Saxena
1c8ab41e8b YARN-5773. RM recovery too slow due to LeafQueue#activateApplications (Bibin A Chundatt via Varun Saxena) 2016-10-29 13:47:39 +05:30
Jason Lowe
1eae719bce YARN-4963. capacity scheduler: Make number of OFF_SWITCH assignments per heartbeat configurable. Contributed by Nathan Roberts 2016-10-28 17:30:15 +00:00
Karthik Kambatla
4df8ed63ed YARN-4743. FairSharePolicy breaks TimSort assumption. (Zephyr Guo and Yufei Gu via kasha) 2016-10-27 17:45:48 -07:00
Subru Krishnan
b2c4f24c31 YARN-2306. Add test for leakage of reservation metrics in fair scheduler. (Hong Zhiguo and Yufei Gu via subru). 2016-10-27 17:43:13 -07:00
Robert Kanter
5877f20f9c HADOOP-10075. Update jetty dependency to version 9 (rkanter) 2016-10-27 16:09:00 -07:00
Subru Krishnan
79ae78dcbe YARN-3568. TestAMRMTokens should use some random port. (Takashi Ohnishi via Subru). 2016-10-27 15:11:12 -07:00