big-data/hadoop
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
25,658 Commits 2 Branches 2 Tags 264 MiB
9e475aede6
Commit Graph

4449 Commits

Author SHA1 Message Date
Viraj Jasani
c0bdba8fac
HADOOP-18017. unguava: remove Preconditions from hadoop-yarn-project modules (#3687) 2021-11-23 13:36:22 +09:00
Szilard Nemeth
7cb887e6c2 YARN-10997. Revisit allocation and reservation logging. Contributed by Andras Gyori 2021-11-12 15:43:45 +01:00
Szilard Nemeth
e220e88eca YARN-10996. Fix race condition of User object acquisitions. Contributed by Andras Gyori 2021-11-12 15:33:39 +01:00
Szilard Nemeth
2e32cc68fc
YARN-10985. Add some tests to verify ACL behaviour in CapacitySchedulerConfiguration (#3570). Contributed by Szilard Nemeth 2021-10-28 15:19:32 +02:00
Szilard Nemeth
d598904046
YARN-10904. Investigate: Remove unnecessary fields from AbstractCSQueue (#3551) contributed by Szilard Nemeth 2021-10-27 19:03:45 +02:00
Szilard Nemeth
66ac476b48
YARN-10924. Clean up CapacityScheduler#initScheduler (#3581) Contributed by Szilard Nemeth 2021-10-27 17:13:49 +02:00
Jack
9cfd8d0a83 YARN-10909. AbstractCSQueue: Annotate all methods with VisibleForTesting that are only used by test code. Contributed by JackWangCS, Szilard Nemeth 2021-10-23 14:47:09 +02:00
9uapaw
32ecaed9c3 YARN-10930. Introduce universal capacity resource vector. Contributed by Andras Gyori 2021-10-22 17:32:33 +02:00
Adam Antal
23772d946b YARN-10948. Rename SchedulerQueue#activeQueue to activateQueue. Contributed by Adam Antal 2021-10-22 16:33:03 +02:00
Ahmed Hussein
d286994009 YARN-1115: Provide optional means for a scheduler to check real user ACLs. Contributed by Eric Payne (epayne) 2021-10-20 22:18:36 +00:00
Szilard Nemeth
20aeb5ecc3
YARN-10916. Investigate and simplify GuaranteedOrZeroCapacityOverTimePolicy#computeQueueManagementChanges. Contributed by Szilard Nemeth 2021-10-20 15:52:37 +02:00
Andras Gyori
35b8441fd9
YARN-10949. Simplify AbstractCSQueue#updateMaxAppRelatedField and find a more meaningful name for this method. Contributed by Andras Gyori 2021-10-20 12:56:41 +02:00
Szilard Nemeth
414d40155c
YARN-10958. Use correct configuration for Group service init in CSMappingPlacementRule (#3560) 
* YARN-10958. Initial commit

* Fix javadoc + behaviour

* Fix review comments

* fix checkstyle + blanks

* fix checkstyle + blanks

* Fix checkstyle + blanks
 2021-10-20 10:48:42 +02:00
9uapaw
616cea2e80 YARN-10954. Remove commented code block from CSQueueUtils#loadCapacitiesByLabelsFromConf. Contributed by Andras Gyori 2021-10-19 13:06:45 +02:00
Szilard Nemeth
025f97c8c2
YARN-10942. Move AbstractCSQueue fields to separate objects that are tracking usage. Contributed by Szilard Nemeth 2021-10-19 12:24:58 +02:00
lujiefsi
ae95caa60e
YARN-10976. Fix resource leak due to Files.walk (#3552) 
Signed-off-by: Akira Ajisaka <aajisaka@apache.org>
 2021-10-18 15:24:15 +09:00
Viraj Jasani
d336227e5c
HADOOP-17963. Replace Guava VisibleForTesting by Hadoop's own annotation in hadoop-yarn-project modules (#3541) 2021-10-14 18:03:01 +09:00
Viraj Jasani
1151edf12e
HADOOP-17956. Replace all default Charset usage with UTF-8 (#3529) 
Signed-off-by: Akira Ajisaka <aajisaka@apache.org>
 2021-10-14 13:07:24 +09:00
Szilard Nemeth
8c271caa81 YARN-6862. Nodemanager resource usage metrics sometimes are negative. Contributed by Benjamin Teke 2021-10-12 17:35:06 +02:00
Benjamin Teke
35eff54556
YARN-10934. Fix LeafQueue#activateApplication NPE when the user of the pending application is missing from usersManager. Contributed by Benjamin Teke 
Co-authored-by: Benjamin Teke <bteke@cloudera.com>
 2021-10-07 20:11:42 +02:00
9uapaw
4b1b6b858a
YARN-10953. Make CapacityScheduler#getOrCreateQueueFromPlacementConte… Contributed by Andras Gyori 2021-10-07 17:09:38 +02:00
Tamas Domok
5535d66fb5
YARN-10692. Do not extend from CapacitySchedulerTestBase when not needed. Contributed by Tamas Domok 
Co-authored-by: Tamas Domok <tdomok@cloudera.com>
 2021-10-07 16:21:05 +02:00
9uapaw
ed8e879320
YARN-10823. Expose all node labels for root without explicit configurations. Contributed by Andras Gyori 2021-10-01 04:20:36 +02:00
Tamas Domok
3c9e3d5321
YARN-10960. Extract test queues and related methods from TestCapacityScheduler. Contributed by Tamas Domok 
Co-authored-by: Tamas Domok <tdomok@cloudera.com>
 2021-10-01 04:05:52 +02:00
Neil
4bd0c36189
YARN-10970. Standby RM should expose prom endpoint (#3480) 
Reviewed-by: Adam Antal <adamantal@apache.org>
Signed-off-by: Akira Ajisaka <aajisaka@apache.org>
 2021-09-29 15:46:55 +09:00
Akshat Bordia
51bd528bd3
YARN-10957. Using invokeConcurrent Overload with Collection in getClusterMetrics (#3439) 2021-09-28 09:51:18 -07:00
Tamas Domok
8f4456d4a1
YARN-10961. TestCapacityScheduler: reuse appHelper where feasible. Contributed by Tamas Domok 
Co-authored-by: Tamas Domok <tdomok@cloudera.com>
 2021-09-21 16:13:04 +02:00
9uapaw
9f6430c9ed
YARN-10897. Introduce QueuePath class. Contributed by Andras Gyori 2021-09-21 16:08:24 +02:00
Szilard Nemeth
4df4389325
YARN-10911. AbstractCSQueue: Create a separate class for usernames and weights that are travelling in a Map. Contributed by Szilard Nemeth 2021-09-20 16:47:46 +02:00
Adam Antal
a9b2469a53
YARN-10950. Code cleanup in QueueCapacities (#3454) 2021-09-19 14:42:02 +02:00
Szilard Nemeth
aa74a303ed
YARN-10913. AbstractCSQueue: Group preemption methods and fields into a separate class (#3420) 2021-09-19 13:11:56 +02:00
Eric Badger
43f0a34dd4 YARN-10935. AM Total Queue Limit goes below per-user AM Limit if parent is full. Contributed by Eric Payne. 2021-09-15 20:03:45 +00:00
Benjamin Teke
5dc2f7b137
YARN-10915. AbstractCSQueue: Simplify complex logic in methods: deriveCapacityFromAbsoluteConfigurations and updateEffectiveResources (#3418) 
Co-authored-by: Benjamin Teke <bteke@cloudera.com>
 2021-09-14 18:05:40 +02:00
Tamas Domok
783d94f5cd
YARN-10917. Investigate and simplify CapacitySchedulerConfigValidator#validateQueueHierarchy (#3403) 
* YARN-10917. Investigate and simplify CapacitySchedulerConfigValidator#validateQueueHierarchy.

Co-authored-by: Tamas Domok <tdomok@cloudera.com>
 2021-09-14 17:54:25 +02:00
Tamas Domok
63c892278f
YARN-10912. AbstractCSQueue#updateConfigurableResourceRequirement: Separate validation logic from initialization logic (#3390) 
- capacityConfigType update is extracted to a separate method
 - validation logic is extracted to a helper function
 - min resource must not be greater than max resource is now checked
   after the max resource is updated

Change-Id: I731c2639281721afed32c30854bafcf048d6ee28

Co-authored-by: Tamas Domok <tdomok@cloudera.com>
 2021-09-14 17:30:44 +02:00
Weihao Zheng
ad1d40970a
YARN-10928. Support default queue config for minimum-user-limit-percent/user-limit-factor (#3389) 
Contributed by Weihao Zheng
 2021-09-13 11:06:53 +08:00
Jack
d8026e387e
YARN-10903. Fix the headroom check in ParentQueue and RegularContainerAllocator for DRF (#3352) 
Contributed by Jie Wang <jie.wang@hulu.com>
 2021-09-13 10:54:11 +08:00
Benjamin Teke
971f1b8b0a
YARN-10872. Replace getPropsWithPrefix calls in AutoCreatedQueueTemplate (#3396) 
Co-authored-by: Benjamin Teke <bteke@cloudera.com>
 2021-09-10 17:32:42 +02:00
9uapaw
811fd23f23
YARN-10852. Optimise CSConfiguration getAllUserWeightsForQueue (#3392) 2021-09-10 16:59:46 +02:00
Benjamin Teke
b229e5a345
YARN-10910. AbstractCSQueue#setupQueueConfigs: Separate validation logic from initialization logic (#3407) 
Co-authored-by: Benjamin Teke <bteke@cloudera.com>
 2021-09-10 16:48:58 +02:00
Tamas Domok
29a6f141d4
YARN-10914. Simplify duplicated code for tracking ResourceUsage in AbstractCSQueue (#3402) 
Co-authored-by: Tamas Domok <tdomok@cloudera.com>
 2021-09-10 15:57:46 +02:00
Akshat Bordia
dee6dc2f89
YARN-10893. Adding metrics for getClusterMetrics and getApplications APIs in FederationClientInterceptor (#3325) 2021-09-09 09:20:57 -07:00
Akshat Bordia
a186460004
YARN-10829. Follow up: Adding null checks before merging ResourceUsage Report (#3252) 2021-09-08 10:36:56 -07:00
Szilard Nemeth
2ff3fc50e4 YARN-10870. Missing user filtering check -> yarn.webapp.filter-entity-list-by-user for RM Scheduler page. Contributed by Gergely Pollak 2021-09-08 18:01:39 +02:00
Jack
4e209a31da
YARN-10919. Remove LeafQueue#scheduler field (#3382) 
Co-authored-by: Jie Wang <jie.wang@hulu.com>
 2021-09-08 16:19:29 +02:00
Benjamin Teke
40e639ad07
YARN-10646. TestCapacitySchedulerWeightMode test descriptor comments doesnt reflect the correct scenario (#3339) 
Co-authored-by: Benjamin Teke <bteke@cloudera.com>
 2021-09-08 16:11:04 +02:00
Swathi Chandrashekar
9512c774c0
YARN-10884: Handle empty owners to parse log files (#3318) 
* YARN 10884 : Parse log files which has empty owner

* Removing the whitespace

* Added a test case for null user

* Fixed indentations

* Fixed the indentation for test cases

Co-authored-by: Swathi C <cswathi@microsoft.com>
 2021-09-07 21:37:51 +05:30
Tamas Domok
16e6030e25
YARN-10891. Extend QueueInfo with max-parallel-apps in CS. (#3314) 
Co-authored-by: Tamas Domok <tdomok@cloudera.com>
 2021-08-27 23:09:54 +02:00
Szilard Nemeth
e06a5cb197 YARN-10838. Implement an optimised version of Configuration getPropsWithPrefix. Contributed by Andras Gyori, Benjamin Teke 2021-08-24 15:27:34 +02:00
srinivasst
4f3f26ce09
YARN-10873: Account for scheduled AM containers before deactivating node (#3287) 
* Account for scheduled AM containers before deactivating node

* Move AM container check to separate method.

* Fix UTs

* Fix UTs

* Remove unnecessary import

* Add timeout for UT
 2021-08-17 14:18:55 +05:30
Tamas Domok
0ba6f35dc2
YARN-10355. Refactor NM ContainerLaunch.java#orderEnvByDependencies 
Co-authored-by: Tamas Domok <tdomok@cloudera.com>
 2021-08-04 15:27:46 +02:00
Tamas Domok
a5811dda7b
YARN-10874. Refactor NM ContainerLaunch#getEnvDependencies's unit tests (#3248) 
Co-authored-by: Tamas Domok <tdomok@cloudera.com>
 2021-08-03 16:19:04 +02:00
Viraj Jasani
ccfa072dc7
HADOOP-17612. Upgrade Zookeeper to 3.6.3 and Curator to 5.2.0 (#3241) 
Signed-off-by: Akira Ajisaka <aajisaka@apache.org>
 2021-08-03 14:44:00 +09:00
zhuqi-lucas
efb3fa2bf5 YARN-10854. Support marking inactive node as untracked without configured include path. Contributed by Tao Yang. 2021-08-02 18:23:33 +08:00
Benjamin Teke
ac0a4e7f58
YARN-10869. CS considers only the default maximum-allocation-mb/vcore property as a maximum when it creates dynamic queues (#3225) 
Co-authored-by: Benjamin Teke <bteke@cloudera.com>
 2021-07-29 17:56:14 +02:00
minni31
d78b300ed4
YARN-10841. Fix token reset synchronization for UAM response token. (#3194) 
YARN-10841. Fix token reset synchronization for UAM response token.  Contributed by Minni Mittal
 2021-07-29 14:55:39 +05:30
Szilard Nemeth
1b9efe58c9 YARN-10790. CS Flexible AQC: Add separate parent and leaf template property. Contributed by Andras Gyori 2021-07-28 16:50:14 +02:00
Szilard Nemeth
8d0297c213 YARN-10727. ParentQueue does not validate the queue on removal. Contributed by Andras Gyori 2021-07-28 14:49:10 +02:00
Szilard Nemeth
fa0289b022 YARN-6221. Entities missing from ATS when summary log file info got returned to the ATS before the domain log. Contributed by Xiaomin Zhang 2021-07-27 16:47:41 +02:00
Benjamin Teke
05b6a1a06a
YARN-10833. Set the X-FRAME-OPTIONS header for the default contexts. (#3203) 
* YARN-10833. Set the X-FRAME-OPTIONS header for the default contexts.

* fixup: YARN-10833. Set the X-FRAME-OPTIONS header for the default contexts.

Co-authored-by: Benjamin Teke <bteke@cloudera.com>
 2021-07-24 11:44:21 +08:00
Akshat Bordia
aa1a5dd413
YARN-10829. Support getApplications API in FederationClientInterceptor (#3135) 
YARN-10829. Support getApplications API in FederationClientInterceptor (#3135)
 2021-07-23 17:07:51 +05:30
zhuqi-lucas
2da9b95d4d YARN-10657. We should make max application per queue to support node label. Contributed by Andras Gyori. 2021-07-22 20:30:43 +08:00
zhuqi-lucas
0441efe1fc YARN-10860. Make max container per heartbeat configs refreshable. Contributed by Eric Badger. 2021-07-21 15:31:44 +08:00
Viraj Jasani
d0ee065cc4
HADOOP-16272. Upgrade HikariCP to 4.0.3 (#3204) 
Signed-off-by: Akira Ajisaka <aajisaka@apache.org>
 2021-07-16 12:17:17 +09:00
Jim Brennan
632f64cadb YARN-10456. RM PartitionQueueMetrics records are named QueueMetrics in Simon metrics registry. Contributed by Eric Payne. 2021-07-15 14:23:31 +00:00
Artem Smotrakov
ba325a8ada
HADOOP-17793. Better token validation (#3189) 
Signed-off-by: Akira Ajisaka <aajisaka@apache.org>
 2021-07-10 13:42:31 +09:00
Viraj Jasani
618c9218ee
HADOOP-17788. Replace IOUtils#closeQuietly usages by Hadoop's own utility (#3171) 
Reviewed-by: Steve Loughran <stevel@apache.org>
Reviewed-by: Akira Ajisaka <aajisaka@apache.org>
Signed-off-by: Takanobu Asanuma <tasanuma@apache.org>
 2021-07-08 16:03:40 +09:00
Jim Brennan
dc6f456e95 YARN-10834. Intra-queue preemption: apps that don't use defined custom resource won't be preempted. Contributed by Eric Payne. 2021-06-28 14:52:19 +00:00
Jim Brennan
7c7d02edbd YARN-10824. Title not set for JHS and NM webpages. Contributed by Bilwa S T. 2021-06-25 20:32:08 +00:00
Peter Bacsko
0934e783cf YARN-10780. Optimise retrieval of configured node labels in CS queues. Contributed by Andras Gyori. 2021-06-24 20:15:10 +02:00
Szilard Nemeth
6562391737 YARN-10813. Set default capacity of root for node labels. Contributed by Andras Gyori 2021-06-16 18:55:09 +02:00
Szilard Nemeth
428478bbe2 YARN-10801. Fix Auto Queue template to properly set all configuration properties. Contributed by Andras Gyori 2021-06-16 18:26:58 +02:00
Szilard Nemeth
e31d06032b YARN-10802. Change Capacity Scheduler minimum-user-limit-percent to accept decimal values. Contributed by Benjamin Teke 2021-06-14 22:33:04 +02:00
Szilard Nemeth
7003997e36 YARN-10789. RM HA startup can fail due to race conditions in ZKConfigurationStore. Contributed by Tarun Parimi 2021-06-12 14:49:52 +02:00
Viraj Jasani
81d7069316
YARN-10805. Replace Guava Lists usage by Hadoop's own Lists in hadoop-yarn-project (#3075) 2021-06-09 15:15:47 +09:00
Viraj Jasani
ca27744770
YARN-10809. Missing dependency causing NoClassDefFoundError in TestHBaseTimelineStorageUtils (#3081) 
Reviewed-by: Takanobu Asanuma <tasanuma@apache.org>
 2021-06-08 19:37:23 -07:00
Prabhu Josephraj
9445abb500 YARN-10792. Set Completed AppAttempt LogsLink to Log Server URL. Contributed by Abhinaba Sarkar 2021-06-08 20:37:40 +05:30
zhuqi-lucas
ec16b1d3b9 YARN-10807. Parents node labels are incorrectly added to child queues in weight mode. Contributed by Benjamin Teke. 2021-06-08 21:03:43 +08:00
Hideyuki Furue
9c7b8cf54e
Fix container-executor 
Signed-off-by: Akira Ajisaka <aajisaka@apache.org>
 2021-06-06 07:05:29 +09:00
Szilard Nemeth
200eec8f2e YARN-10796. Capacity Scheduler: dynamic queue cannot scale out properly if its capacity is 0%. Contributed by Peter Bacsko 2021-06-03 13:27:36 +02:00
Szilard Nemeth
2707f69251 YARN-10787. Queue submit ACL check is wrong when CS queue is ambiguous. Contributed by Gergely Pollak 2021-06-01 16:01:39 +02:00
Gergely Pollak
e9339aa376 YARN-10797. Logging parameter issues in scheduler package. Contributed by Szilard Nemeth 2021-06-01 15:57:22 +02:00
Szilard Nemeth
b86a6eb871 YARN-10782. Extend /scheduler endpoint with template properties. Contributed by Andras Gyori 2021-05-25 18:27:53 +02:00
Szilard Nemeth
2541efa496 YARN-10783. Allow definition of auto queue template properties in root. Contributed by Andras Gyori 2021-05-25 13:55:59 +02:00
Viraj Jasani
996d31f2dc
HADOOP-17721. Replace Guava Sets usage by Hadoop's own Sets in hadoop-yarn-project (#3033) 
Signed-off-by: Takanobu Asanuma <tasanuma@apache.org>
 2021-05-25 18:10:20 +09:00
zhuqi-lucas
59172ada90 YARN-10771. Add cluster metric for size of SchedulerEventQueue and RMEventQueue. Contributed by chaosju. 2021-05-24 23:12:07 +08:00
Szilard Nemeth
1e44bdb84c YARN-7769. FS QueueManager should not create default queue at init. Contributed by Benjamin Teke 2021-05-22 14:55:01 +02:00
Peter Bacsko
eb72628e15 YARN-10258. Add metrics for 'ApplicationsRunning' in NodeManager. Contributed by ANANDA G B. 2021-05-17 21:07:29 +02:00
Peter Bacsko
8891e5c028 YARN-10763. Add the number of containers assigned per second metrics to ClusterMetrics. Contributed by chaosju. 2021-05-17 13:30:12 +02:00
lujiefsi
d92a25b790
YARN-10555. Missing access check before getAppAttempts (#2608) 
Co-authored-by: lujie <lujie@foxmail.com>
Signed-off-by: Akira Ajisaka <aajisaka@apache.org>
 2021-05-17 13:44:38 +09:00
zhuqi
e7f0e8073b YARN-10761: Add more event type to RM Dispatcher event metrics. Contributed by Qi Zhu. 2021-05-14 13:36:07 +08:00
zhuqi
d2b0675d61
YARN-10737: Fix typos in CapacityScheduler#schedule. (#2911) 
Contributed by Qi Zhu.
 2021-05-14 13:12:28 +08:00
Peter Bacsko
626be24c3e YARN-10571. Refactor dynamic queue handling logic. Contributed by Andras Gyori. 2021-05-12 14:54:47 +02:00
Akira Ajisaka
7cb7b8adc1
YARN-10756. Remove additional junit 4.11 dependency from javadoc. (#2960) 
Reviewed-by: Viraj Jasani <vjasani@apache.org>
Reviewed-by: Bilwa S T <stbilwa@gmail.com>
Signed-off-by: Takanobu Asanuma <tasanuma@apache.org>
 2021-05-06 23:37:45 +09:00
Wei-Chiu Chuang
f1e1809029
HADOOP-17653. Do not use guava's Files.createTempDir(). (#2945) 
Reviewed-by: Steve Loughran <stevel@apache.org>
Signed-off-by: Akira Ajisaka <aajisaka@apache.org>
 2021-05-02 11:10:32 +09:00
Eric Badger
803ac4b1a0 YARN-10707. Support custom resources in ResourceUtilization, and update Node GPU Utilization to use. Contributed by Qi Zhu 2021-04-29 17:42:13 +00:00
Peter Bacsko
9166bfeb74 YARN-10637. fs2cs: add queue autorefresh policy during conversion. Contributed by Qi Zhu. 2021-04-24 10:10:10 +02:00
Eric Badger
6857a05d6a YARN-10479. Can't remove all node labels after add node label without 
nodemanager port, broken by YARN-10647. Contributed by D M Murali Krishna Reddy
 2021-04-23 22:02:04 +00:00
Jim Brennan
1cbe35946b YARN-10743. Add a policy for not aggregating for containers which are killed because exceeding container log size limit. Contributed by Qi Zhu. 2021-04-23 21:35:09 +00:00
Peter Bacsko
14a84c47b0 YARN-10705. Misleading DEBUG log for container assignment needs to be removed when the container is actually reserved, not assigned in FairScheduler. Contributed by Siddharth Ahuja. 2021-04-23 17:38:47 +02:00