Commit Graph

830 Commits

Author SHA1 Message Date
Jian He c53420f583 YARN-2637. Fixed max-am-resource-percent calculation in CapacityScheduler when activating applications. Contributed by Craig Welch 2015-01-13 17:32:07 -08:00
Robert Kanter 51881535e6 YARN-2643. Don't create a new DominantResourceCalculator on every FairScheduler.allocate call. (kasha via rkanter) 2015-01-12 17:51:00 -08:00
Robert Kanter ae7bf31fe1 YARN-3027. Scheduler should use totalAvailable resource from node instead of availableResource for maxAllocation. (adhoot via rkanter) 2015-01-12 10:47:52 -08:00
Jian He a260406268 YARN-3014. Replaces labels on a host should update all NM's labels on that host. Contributed by Wangda Tan 2015-01-09 17:49:53 -08:00
Zhijie Shen dc2eaa26b2 YARN-2996. Improved synchronization and I/O operations of FS- and Mem- RMStateStore. Contributed by Yi Liu. 2015-01-08 09:47:02 -08:00
Jian He e13a484a2b YARN-3010. Fixed findbugs warning in AbstractYarnScheduler. Contributed by Yi Liu 2015-01-07 14:12:20 -08:00
Zhijie Shen 60103fca04 YARN-2427. Added the API of moving apps between queues in RM web services. Contributed by Varun Vasudev. 2015-01-06 14:37:44 -08:00
Jian He dd57c2047b YARN-2978. Fixed potential NPE while getting queue info. Contributed by Varun Saxena 2015-01-06 14:15:57 -08:00
Karthik Kambatla 0c4b112677 YARN-2881. [YARN-2574] Implement PlanFollower for FairScheduler. (Anubhav Dhoot via kasha) 2015-01-06 04:42:10 +05:30
Zhijie Shen 562a701945 YARN-2958. Made RMStateStore not update the last sequence number when updating the delegation token. Contributed by Varun Saxena. 2015-01-05 13:33:07 -08:00
Tsuyoshi Ozawa ddc5be48fc YARN-2922. ConcurrentModificationException in CapacityScheduler's LeafQueue. Contributed by Rohith Sharmaks. 2015-01-05 00:08:31 +09:00
Karthik Kambatla e7257acd8a YARN-2998. Abstract out scheduler independent PlanFollower components. (Anubhav Dhoot via kasha) 2014-12-30 19:55:24 -08:00
Jian He e2351c7ae2 YARN-2987. Fixed ClientRMService#getQueueInfo to check against queue and app ACLs. Contributed by Varun Saxena 2014-12-30 17:15:37 -08:00
Jian He b7442bf92e YARN-2493. Added node-labels page on RM web UI. Contributed by Wangda Tan 2014-12-30 16:49:01 -08:00
Jian He 746ad6e989 Revert "YARN-2492(wrong jira number). Added node-labels page on RM web UI. Contributed by Wangda Tan"
This reverts commit 5f57b904f5.
2014-12-30 16:48:49 -08:00
Jian He 5f57b904f5 YARN-2492. Added node-labels page on RM web UI. Contributed by Wangda Tan 2014-12-30 15:38:28 -08:00
Zhijie Shen 241d3b3a50 YARN-2938. Fixed new findbugs warnings in hadoop-yarn-resourcemanager and hadoop-yarn-applicationhistoryservice. Contributed by Varun Saxena. 2014-12-29 09:59:35 -08:00
Jian He 1454efe5d4 YARN-2992. ZKRMStateStore crashes due to session expiry. Contributed by Karthik Kambatla 2014-12-26 17:59:58 -08:00
Junping Du 40ee4bff65 YARN-2993. Several fixes (missing acl check, error log msg ...) and some refinement in AdminService. (Contributed by Yi Liu) 2014-12-26 06:59:25 -08:00
Jian He 4f18018b7a YARN-2946. Fixed potential deadlock in RMStateStore. Contributed by Rohith Sharmaks 2014-12-23 22:14:29 -08:00
Jian He 149512a837 YARN-2837. Support TimeLine server to recover delegation token when restarting. Contributed by Zhijie Shen 2014-12-23 18:25:37 -08:00
Jian He 0d89859b51 YARN-2340. Fixed NPE when queue is stopped during RM restart. Contributed by Rohith Sharmaks 2014-12-22 21:53:22 -08:00
Jian He fdf042dfff YARN-2920. Changed CapacityScheduler to kill containers on nodes where node labels are changed. Contributed by Wangda Tan 2014-12-22 16:51:15 -08:00
Junping Du a696fbb001 YARN-2939. Fix new findbugs warnings in hadoop-yarn-common. (Li Lu via junping_du) 2014-12-22 03:06:37 -08:00
Karthik Kambatla 24ee9e3431 YARN-2975. FSLeafQueue app lists are accessed without required locks. (kasha) 2014-12-20 12:17:50 -08:00
Jian He 808cba3821 YARN-2952. Fixed incorrect version check in StateStore. Contributed by Rohith Sharmaks 2014-12-19 16:56:30 -08:00
Karthik Kambatla a22ffc3188 YARN-2738. [YARN-2574] Add FairReservationSystem for FairScheduler. (Anubhav Dhoot via kasha) 2014-12-19 15:37:12 -08:00
Jason Lowe 0402bada19 YARN-2964. RM prematurely cancels tokens for jobs that submit jobs (oozie). Contributed by Jian He 2014-12-18 23:28:18 +00:00
Junping Du 2b4b0e8847 YARN-2972. DelegationTokenRenewer thread pool never expands. Contributed by Jason Lowe 2014-12-17 17:40:00 -08:00
Karthik Kambatla f2d150ea12 YARN-2964. FSLeafQueue#assignContainer - document the reason for using both write and read locks. (Tsuyoshi Ozawa via kasha) 2014-12-17 12:03:16 -08:00
Steve Loughran 3681de2039 YARN-2912 Jersey Tests failing with port in use. (varun saxena via stevel) 2014-12-12 17:11:07 +00:00
Devaraj K bda748ac3a YARN-2243. Order of arguments for Preconditions.checkNotNull() is wrong in
SchedulerApplicationAttempt ctor. Contributed by Devaraj K.
2014-12-12 12:34:43 +05:30
Jian He 2ed90a57fd YARN-2930. Fixed TestRMRestart#testRMRestartRecoveringNodeLabelManager intermittent failure. Contributed by Wangda Tan 2014-12-09 16:48:04 -08:00
Karthik Kambatla a2e07a5456 YARN-2910. FSLeafQueue can throw ConcurrentModificationException. (Wilfred Spiegelenburg via kasha) 2014-12-09 14:00:31 -08:00
Jian He e69af836f3 YARN-2869. CapacityScheduler should trim sub queue names when parse configuration. Contributed by Wangda Tan 2014-12-05 17:33:39 -08:00
Jason Lowe 4b13082199 YARN-2056. Disable preemption at Queue level. Contributed by Eric Payne 2014-12-05 21:06:48 +00:00
Jian He 258623ff8b YARN-2301. Improved yarn container command. Contributed by Naganarasimha G R 2014-12-04 12:53:18 -08:00
Jian He 73fbb3c66b YARN-2880. Added a test to make sure node labels will be recovered if RM restart is enabled. Contributed by Rohith Sharmaks 2014-12-03 17:14:52 -08:00
Karthik Kambatla 799353e2c7 YARN-2874. Dead lock in DelegationTokenRenewer which blocks RM to execute any further apps. (Naganarasimha G R via kasha) 2014-12-03 13:44:41 -08:00
Jian He 392c3aaea8 YARN-2894. Fixed a bug regarding application view acl when RM fails over. Contributed by Rohith Sharmaks 2014-12-02 17:16:35 -08:00
Jian He 52bcefca8b YARN-2136. Changed RMStateStore to ignore store opearations when fenced. Contributed by Varun Saxena 2014-12-02 10:54:48 -08:00
Jian He a7fba0bc28 YARN-2765. Added leveldb-based implementation for RMStateStore. Contributed by Jason Lowe 2014-12-01 16:38:25 -08:00
Junping Du c732ed760e YARN-2907. SchedulerNode#toString should print all resource detail instead of only memory. (Contributed by Rohith) 2014-12-01 05:38:22 -08:00
Jian He 8a7ca13b13 YARN-2906. CapacitySchedulerPage shows HTML tags for a queue's Active Users. Contributed by Jason Lowe 2014-11-25 16:16:56 -08:00
Jian He 5805a81efb YARN-2404. Removed ApplicationAttemptState and ApplicationState class in RMStateStore. Contributed by Tsuyoshi OZAWA 2014-11-25 12:48:22 -08:00
Zhijie Shen e37a4ff0c1 YARN-2697. Remove useless RMAuthenticationHandler. Contributed by Haosong Huang. 2014-11-24 14:58:59 -08:00
Sandy Ryza a128cca305 YARN-2669. FairScheduler: queue names shouldn't allow periods (Wei Yan via Sandy Ryza) 2014-11-21 16:06:41 -08:00
Karthik Kambatla 3114d4731d YARN-2604. Scheduler should consider max-allocation-* in conjunction with the largest node. (Robert Kanter via kasha) 2014-11-21 10:32:28 -08:00
Karthik Kambatla a9a0cc3679 YARN-2315. FairScheduler: Set current capacity in addition to capacity. (Zhihai Xu via kasha) 2014-11-19 20:15:40 -08:00
Karthik Kambatla c90fb84aaa YARN-2802. ClusterMetrics to include AM launch and register delays. (Zhihai Xu via kasha) 2014-11-19 19:50:12 -08:00
Jian He 9cb8b75ba5 YARN-2865. Fixed RM to always create a new RMContext when transtions from StandBy to Active. Contributed by Rohith Sharmaks 2014-11-19 19:48:52 -08:00
Karthik Kambatla 2fce6d6141 YARN-2690. [YARN-2574] Make ReservationSystem and its dependent classes independent of Scheduler type. (Anubhav Dhoot via kasha) 2014-11-17 16:45:57 -08:00
Jason Lowe 81c9d17af8 YARN-2414. RM web UI: app page will crash if app is failed before any attempt has been created. Contributed by Wangda Tan 2014-11-17 21:15:48 +00:00
Jian He 9b860661a5 YARN-2432. RMStateStore should process the pending events before close. Contributed by Varun Saxena 2014-11-14 17:24:09 -08:00
Sandy Ryza 1a47f890ba YARN-2811. In Fair Scheduler, reservation fulfillments shouldn't ignore max share (Siqi Li via Sandy Ryza) 2014-11-14 15:18:56 -08:00
Jian He d005404ef7 YARN-2856. Fixed RMAppImpl to handle ATTEMPT_KILLED event at ACCEPTED state on app recovery. Contributed by Rohith Sharmaks 2014-11-13 15:34:26 -08:00
Vinod Kumar Vavilapalli 3651fe1b08 YARN-2853. Fixed a bug in ResourceManager causing apps to hang when the user kill request races with ApplicationMaster finish. Contributed by Jian He. 2014-11-13 08:13:03 -08:00
Jason Lowe f8aefa5e9c YARN-2780. Log aggregated resource allocation in rm-appsummary.log. Contributed by Eric Payne 2014-11-12 17:01:15 +00:00
Vinod Kumar Vavilapalli e76faebc95 YARN-2834. Fixed ResourceManager to ignore token-renewal failures on recovery consistent with the (somewhat incorrect) behaviour in the non-recovery case. Contributed by Jian He. 2014-11-09 18:56:58 -08:00
Arun C. Murthy 43cd07b408 YARN-2830. Add backwords compatible ContainerId.newInstance constructor. Contributed by Jonathan Eagles. 2014-11-09 14:57:37 -08:00
Zhijie Shen 9a4e0d343e YARN-2505. Supported get/add/remove/change labels in RM REST API. Contributed by Craig Welch. 2014-11-07 20:35:46 -08:00
Vinod Kumar Vavilapalli df36edf751 YARN-2826. Fixed user-groups mappings' refresh bug caused by YARN-2826. Contributed by Wangda Tan. 2014-11-07 19:43:55 -08:00
Vinod Kumar Vavilapalli 4cfd5bc7c1 YARN-2753. Fixed a bunch of bugs in the NodeLabelsManager classes. Contributed by Zhihai xu. 2014-11-07 14:15:53 -08:00
Vinod Kumar Vavilapalli 2ac1be7dec YARN-2824. Fixed Capacity Scheduler to not crash when some node-labels are not mapped to queues by making default capacities per label to be zero. Contributed by Wangda Tan. 2014-11-07 10:39:37 -08:00
Xuan 1e97f2f094 YARN-2810. TestRMProxyUsersConf fails on Windows VMs. Contributed by Varun Vasudev 2014-11-07 09:44:43 -08:00
Vinod Kumar Vavilapalli a5657182a7 YARN-2823. Fixed ResourceManager app-attempt state machine to inform schedulers about previous finished attempts of a running appliation to avoid expectation mismatch w.r.t transferred containers. Contributed by Jian He. 2014-11-07 09:28:36 -08:00
Vinod Kumar Vavilapalli a3839a9fbf YARN-2744. Fixed CapacityScheduler to validate node-labels correctly against queues. Contributed by Wangda Tan. 2014-11-06 17:28:12 -08:00
Jian He 395275af86 YARN-2579. Fixed a deadlock issue when EmbeddedElectorService and FatalEventDispatcher try to transition RM to StandBy at the same time. Contributed by Rohith Sharmaks 2014-11-05 16:59:54 -08:00
Vinod Kumar Vavilapalli 834e931d8e YARN-2805. Fixed ResourceManager to load HA configs correctly before kerberos login. Contributed by Wangda Tan. 2014-11-05 15:30:33 -08:00
Zhijie Shen b4c951ab83 YARN-2767. Added a test case to verify that http static user cannot kill or submit apps in the secure mode. Contributed by Varun Vasudev. 2014-11-05 10:57:38 -08:00
Karthik Kambatla b2cd269802 YARN-2010. Handle app-recovery failures gracefully. (Jian He and Karthik Kambatla via kasha) 2014-11-04 17:45:24 -08:00
Vinod Kumar Vavilapalli ec6cbece8e YARN-2795. Fixed ResourceManager to not crash loading node-label data from HDFS in secure mode. Contributed by Wangda Tan. 2014-11-03 13:44:06 -08:00
Zhijie Shen 27715ec63b YARN-2785. Fixed intermittent TestContainerResourceUsage failure. Contributed by Varun Vasudev. 2014-11-02 15:20:40 -08:00
Vinod Kumar Vavilapalli d1828d9443 YARN-2779. Fixed ResourceManager to not require delegation tokens for communicating with Timeline Service. Contributed by Zhijie Shen. 2014-10-30 23:16:19 -07:00
Vinod Kumar Vavilapalli e0233c16eb YARN-2698. Moved some node label APIs to be correctly placed in client protocol. Contributed by Wangda Tan. 2014-10-30 22:59:31 -07:00
Karthik Kambatla 179cab81e0 YARN-2712. TestWorkPreservingRMRestart: Augment FS tests with queue and headroom checks. (Tsuyoshi Ozawa via kasha) 2014-10-30 00:29:07 -07:00
Karthik Kambatla 782971ae7a YARN-2742. FairSchedulerConfiguration should allow extra spaces between value and unit. (Wei Yan via kasha) 2014-10-29 10:24:57 -07:00
Jian He d5e0a09721 YARN-2503. Added node lablels in web UI. Contributed by Wangda Tan 2014-10-28 17:57:54 -07:00
Vinod Kumar Vavilapalli a16d022ca4 YARN-2704. Changed ResourceManager to optionally obtain tokens itself for the sake of localization and log-aggregation for long-running services. Contributed by Jian He. 2014-10-27 15:49:47 -07:00
Xuan ce1a4419a6 YARN-2726. CapacityScheduler should explicitly log when an accessible label has no capacity. Contributed by Wangda Tan 2014-10-27 10:26:21 -07:00
Vinod Kumar Vavilapalli 0186645505 YARN-2743. Fixed a bug in ResourceManager that was causing RMDelegationToken identifiers to be tampered and thus causing app submission failures in secure mode. Contributed by Jian He. 2014-10-26 11:14:34 -07:00
Jian He 5864dd99a4 YARN-1915. Fixed a race condition that client could use the ClientToAMToken to contact with AM before AM actually receives the ClientToAMTokenMasterKey. Contributed by Jason Lowe 2014-10-24 22:47:56 -07:00
Zhijie Shen 0f3b6900be YARN-2209. Replaced AM resync/shutdown command with corresponding exceptions and made related MR changes. Contributed by Jian He. 2014-10-23 21:56:03 -07:00
Zhijie Shen 7e3b5e6f5c YARN-2721. Suppress NodeExist exception thrown by ZKRMStateStore when it retries creating znode. Contributed by Jian He. 2014-10-21 21:49:17 -07:00
Vinod Kumar Vavilapalli c0e034336c YARN-2715. Fixed ResourceManager to respect common configurations for proxy users/groups beyond just the YARN level config. Contributed by Zhijie Shen. 2014-10-21 20:09:40 -07:00
Vinod Kumar Vavilapalli 8256766498 YARN-2504. Enhanced RM Admin CLI to support management of node-labels. Contribyted by Wangda Tan. 2014-10-18 12:07:40 -07:00
Vinod Kumar Vavilapalli 39063cd36f YARN-2676. Enhanced Timeline auth-filter to support proxy users. Contributed by Zhijie Shen. 2014-10-17 22:02:50 -07:00
Vinod Kumar Vavilapalli e9c66e8fd2 YARN-2705. Fixed bugs in ResourceManager node-label manager that were causing test-failures: added a dummy in-memory labels-manager. Contributed by Wangda Tan. 2014-10-17 18:26:12 -07:00
Jian He c3de2412eb YARN-1879. Marked Idempotent/AtMostOnce annotations to ApplicationMasterProtocol for RM fail over. Contributed by Tsuyoshi OZAWA 2014-10-17 16:35:27 -07:00
Jian He a6aa6e42ca YARN-2588. Standby RM fails to transitionToActive if previous transitionToActive failed with ZK exception. Contributed by Rohith Sharmaks 2014-10-17 10:54:24 -07:00
Vinod Kumar Vavilapalli abae63caf9 YARN-2699. Fixed a bug in CommonNodeLabelsManager that caused tests to fail when using ephemeral ports on NodeIDs. Contributed by Wangda Tan. 2014-10-17 08:58:08 -07:00
Vinod Kumar Vavilapalli b3056c266a YARN-2685. Fixed a bug in CommonNodeLabelsManager that caused wrong resource tracking per label when a host runs multiple node-managers. Contributed by Wangda Tan. 2014-10-15 18:47:26 -07:00
Vinod Kumar Vavilapalli f2ea555ac6 YARN-2496. Enhanced Capacity Scheduler to have basic support for allocating resources based on node-labels. Contributed by Wangda Tan.
YARN-2500. Ehnaced ResourceManager to support schedulers allocating resources based on node-labels. Contributed by Wangda Tan.
2014-10-15 18:33:06 -07:00
Jian He 0af1a2b5bc YARN-2312. Deprecated old ContainerId#getId API and updated MapReduce to use ContainerId#getContainerId instead. Contributed by Tsuyoshi OZAWA 2014-10-15 15:22:07 -07:00
Zhijie Shen 1220bb72d4 YARN-2656. Made RM web services authentication filter support proxy user. Contributed by Varun Vasudev and Zhijie Shen. 2014-10-14 21:50:46 -07:00
Zhijie Shen cdce88376a HADOOP-11181. Generalized o.a.h.s.t.d.DelegationTokenManager to handle all sub-classes of AbstractDelegationTokenIdentifier. Contributed by Zhijie Shen. 2014-10-14 11:35:38 -07:00
Karthik Kambatla da709a2eac YARN-2641. Decommission nodes on -refreshNodes instead of next NM-RM heartbeat. (Zhihai Xu via kasha) 2014-10-13 16:23:04 -07:00
Jian He f9680d9a16 YARN-2308. Changed CapacityScheduler to explicitly throw exception if the queue
to which the apps were submitted is changed across RM restart. Contributed by Craig Welch & Chang Li
2014-10-13 14:09:04 -07:00
Zhijie Shen 4aed2d8e91 YARN-2651. Spun off LogRollingInterval from LogAggregationContext. Contributed by Xuan Gong. 2014-10-13 10:54:09 -07:00
Vinod Kumar Vavilapalli db7f165319 YARN-2494. Added NodeLabels Manager internal API and implementation. Contributed by Wangda Tan. 2014-10-10 11:44:21 -07:00