[jira] [Created] (SPARK-4585) Spark dynamic scaling executors use upper limit value as default.
Chengxiang Li created SPARK-4585: Summary: Spark dynamic scaling executors use upper limit value as default. Key: SPARK-4585 URL: https://issues.apache.org/jira/browse/SPARK-4585 Project: Spark Issue Type: Bug Components: Spark Core, YARN Affects Versions: 1.1.0 Reporter: Chengxiang Li With SPARK-3174, one can configure a minimum and maximum number of executors for a Spark application on Yarn. However, the application always starts with the maximum. It seems more reasonable, at least for Hive on Spark, to start from the minimum and scale up as needed up to the maximum. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Created] (SPARK-4955) Executor does not get killed after configured interval.
Chengxiang Li created SPARK-4955: Summary: Executor does not get killed after configured interval. Key: SPARK-4955 URL: https://issues.apache.org/jira/browse/SPARK-4955 Project: Spark Issue Type: Bug Components: YARN Affects Versions: 1.2.0 Reporter: Chengxiang Li With executor dynamic scaling enabled, in yarn-cluster mode, after query finished and spark.dynamicAllocation.executorIdleTimeout interval, executor number is not reduced to configured min number. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-4955) Executor does not get killed after configured interval.
[ https://issues.apache.org/jira/browse/SPARK-4955?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14258138#comment-14258138 ] Chengxiang Li commented on SPARK-4955: -- I verified this feature with Hive on Spark, it works well in yarn-client mode, while failed to reduce to min executor number after idle timeout in yarn-cluster mode, here is the related appmaster log: {noformat} 14/12/24 16:12:50 INFO cluster.YarnClusterSchedulerBackend: Requesting to kill executor(s) 2 14/12/24 16:12:50 WARN cluster.YarnClusterSchedulerBackend: Attempted to kill executors before the AM has registered! 14/12/24 16:12:50 WARN spark.ExecutorAllocationManager: Unable to reach the cluster manager to kill executor 2! 14/12/24 16:12:50 INFO cluster.YarnClusterSchedulerBackend: Requesting to kill executor(s) 1 14/12/24 16:12:50 WARN cluster.YarnClusterSchedulerBackend: Attempted to kill executors before the AM has registered! 14/12/24 16:12:50 WARN spark.ExecutorAllocationManager: Unable to reach the cluster manager to kill executor 1! 14/12/24 16:12:50 INFO cluster.YarnClusterSchedulerBackend: Requesting to kill executor(s) 4 14/12/24 16:12:50 WARN cluster.YarnClusterSchedulerBackend: Attempted to kill executors before the AM has registered! 14/12/24 16:12:50 WARN spark.ExecutorAllocationManager: Unable to reach the cluster manager to kill executor 4! 14/12/24 16:12:50 INFO cluster.YarnClusterSchedulerBackend: Requesting to kill executor(s) 3 14/12/24 16:12:50 WARN cluster.YarnClusterSchedulerBackend: Attempted to kill executors before the AM has registered! 14/12/24 16:12:50 WARN spark.ExecutorAllocationManager: Unable to reach the cluster manager to kill executor 3! {noformat} > Executor does not get killed after configured interval. > --- > > Key: SPARK-4955 > URL: https://issues.apache.org/jira/browse/SPARK-4955 > Project: Spark > Issue Type: Bug > Components: YARN >Affects Versions: 1.2.0 >Reporter: Chengxiang Li > > With executor dynamic scaling enabled, in yarn-cluster mode, after query > finished and spark.dynamicAllocation.executorIdleTimeout interval, executor > number is not reduced to configured min number. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-4955) Executor does not get killed after configured interval.
[ https://issues.apache.org/jira/browse/SPARK-4955?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14258139#comment-14258139 ] Chengxiang Li commented on SPARK-4955: -- cc:[~andrewor14] > Executor does not get killed after configured interval. > --- > > Key: SPARK-4955 > URL: https://issues.apache.org/jira/browse/SPARK-4955 > Project: Spark > Issue Type: Bug > Components: YARN >Affects Versions: 1.2.0 >Reporter: Chengxiang Li > > With executor dynamic scaling enabled, in yarn-cluster mode, after query > finished and spark.dynamicAllocation.executorIdleTimeout interval, executor > number is not reduced to configured min number. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Created] (SPARK-3199) native Java spark listener API support
Chengxiang Li created SPARK-3199: Summary: native Java spark listener API support Key: SPARK-3199 URL: https://issues.apache.org/jira/browse/SPARK-3199 Project: Spark Issue Type: New Feature Components: Java API Reporter: Chengxiang Li Current spark listener API is totally scala style, full of case classes and scala collections, a native Java spark listener API would be much friendly for Java users. -- This message was sent by Atlassian JIRA (v6.2#6252) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-2633) enhance spark listener API to gather more spark job information
[ https://issues.apache.org/jira/browse/SPARK-2633?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Chengxiang Li updated SPARK-2633: - Summary: enhance spark listener API to gather more spark job information (was: support register spark listener to listener bus with Java API) > enhance spark listener API to gather more spark job information > --- > > Key: SPARK-2633 > URL: https://issues.apache.org/jira/browse/SPARK-2633 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li >Priority: Critical > Labels: hive > Attachments: Spark listener enhancement for Hive on Spark job monitor > and statistic.docx > > > Currently user can only register spark listener with Scala API, we should add > this feature to Java API as well. -- This message was sent by Atlassian JIRA (v6.2#6252) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-2633) enhance spark listener API to gather more spark job information
[ https://issues.apache.org/jira/browse/SPARK-2633?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Chengxiang Li updated SPARK-2633: - Description: Based on Hive on Spark job status monitoring and statistic collection requirement, try to enhance spark listener API to gather more spark job information. (was: Currently user can only register spark listener with Scala API, we should add this feature to Java API as well.) > enhance spark listener API to gather more spark job information > --- > > Key: SPARK-2633 > URL: https://issues.apache.org/jira/browse/SPARK-2633 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li >Priority: Critical > Labels: hive > Attachments: Spark listener enhancement for Hive on Spark job monitor > and statistic.docx > > > Based on Hive on Spark job status monitoring and statistic collection > requirement, try to enhance spark listener API to gather more spark job > information. -- This message was sent by Atlassian JIRA (v6.2#6252) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-2633) enhance spark listener API to gather more spark job information
[ https://issues.apache.org/jira/browse/SPARK-2633?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14108887#comment-14108887 ] Chengxiang Li commented on SPARK-2633: -- I would start to work on this issue, for better isolation, this JIRA would focus on spark listener API enhancement, and I've created SPARK-3199 to track native Spark listener API implementation. > enhance spark listener API to gather more spark job information > --- > > Key: SPARK-2633 > URL: https://issues.apache.org/jira/browse/SPARK-2633 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li >Priority: Critical > Labels: hive > Attachments: Spark listener enhancement for Hive on Spark job monitor > and statistic.docx > > > Based on Hive on Spark job status monitoring and statistic collection > requirement, try to enhance spark listener API to gather more spark job > information. -- This message was sent by Atlassian JIRA (v6.2#6252) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-2633) enhance spark listener API to gather more spark job information
[ https://issues.apache.org/jira/browse/SPARK-2633?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14113611#comment-14113611 ] Chengxiang Li commented on SPARK-2633: -- It's quite subjective I think, like Hive on MR display job progress by task finished percentage, while Hive on Tez display job progress with exact running/failed/finished task number. I think it's better we collect more detail job status info while it does not introduce much extra effort. > enhance spark listener API to gather more spark job information > --- > > Key: SPARK-2633 > URL: https://issues.apache.org/jira/browse/SPARK-2633 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li >Priority: Critical > Labels: hive > Attachments: Spark listener enhancement for Hive on Spark job monitor > and statistic.docx > > > Based on Hive on Spark job status monitoring and statistic collection > requirement, try to enhance spark listener API to gather more spark job > information. -- This message was sent by Atlassian JIRA (v6.2#6252) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-2633) enhance spark listener API to gather more spark job information
[ https://issues.apache.org/jira/browse/SPARK-2633?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Chengxiang Li updated SPARK-2633: - Attachment: (was: Spark listener enhancement for Hive on Spark job monitor and statistic.docx) > enhance spark listener API to gather more spark job information > --- > > Key: SPARK-2633 > URL: https://issues.apache.org/jira/browse/SPARK-2633 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li >Priority: Critical > Labels: hive > Attachments: Spark listener enhancement for Hive on Spark job monitor > and statistic.docx > > > Based on Hive on Spark job status monitoring and statistic collection > requirement, try to enhance spark listener API to gather more spark job > information. -- This message was sent by Atlassian JIRA (v6.2#6252) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-2633) enhance spark listener API to gather more spark job information
[ https://issues.apache.org/jira/browse/SPARK-2633?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Chengxiang Li updated SPARK-2633: - Attachment: Spark listener enhancement for Hive on Spark job monitor and statistic.docx > enhance spark listener API to gather more spark job information > --- > > Key: SPARK-2633 > URL: https://issues.apache.org/jira/browse/SPARK-2633 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li >Priority: Critical > Labels: hive > Attachments: Spark listener enhancement for Hive on Spark job monitor > and statistic.docx > > > Based on Hive on Spark job status monitoring and statistic collection > requirement, try to enhance spark listener API to gather more spark job > information. -- This message was sent by Atlassian JIRA (v6.2#6252) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-2895) Support mapPartitionsWithContext in Spark Java API
[ https://issues.apache.org/jira/browse/SPARK-2895?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14116221#comment-14116221 ] Chengxiang Li commented on SPARK-2895: -- pull request:[https://github.com/apache/spark/pull/2194] > Support mapPartitionsWithContext in Spark Java API > -- > > Key: SPARK-2895 > URL: https://issues.apache.org/jira/browse/SPARK-2895 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li > Labels: hive > > This is a requirement from Hive on Spark, mapPartitionsWithContext only > exists in Spark Scala API, we expect to access from Spark Java API. -- This message was sent by Atlassian JIRA (v6.2#6252) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-2895) Support mapPartitionsWithContext in Spark Java API
[ https://issues.apache.org/jira/browse/SPARK-2895?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Chengxiang Li updated SPARK-2895: - Description: This is a requirement from Hive on Spark, mapPartitionsWithContext only exists in Spark Scala API, we expect to access from Spark Java API. For HIVE-7627, HIVE-7843, Hive operators which are invoked in mapPartitions closure need to get taskId. was:This is a requirement from Hive on Spark, mapPartitionsWithContext only exists in Spark Scala API, we expect to access from Spark Java API. > Support mapPartitionsWithContext in Spark Java API > -- > > Key: SPARK-2895 > URL: https://issues.apache.org/jira/browse/SPARK-2895 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li >Assignee: Chengxiang Li > Labels: hive > > This is a requirement from Hive on Spark, mapPartitionsWithContext only > exists in Spark Scala API, we expect to access from Spark Java API. > For HIVE-7627, HIVE-7843, Hive operators which are invoked in mapPartitions > closure need to get taskId. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-2633) enhance spark listener API to gather more spark job information
[ https://issues.apache.org/jira/browse/SPARK-2633?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Chengxiang Li updated SPARK-2633: - Attachment: Spark listener enhancement for Hive on Spark job monitor and statistic.docx > enhance spark listener API to gather more spark job information > --- > > Key: SPARK-2633 > URL: https://issues.apache.org/jira/browse/SPARK-2633 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li >Priority: Critical > Labels: hive > Attachments: Spark listener enhancement for Hive on Spark job monitor > and statistic.docx > > > Based on Hive on Spark job status monitoring and statistic collection > requirement, try to enhance spark listener API to gather more spark job > information. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-2633) enhance spark listener API to gather more spark job information
[ https://issues.apache.org/jira/browse/SPARK-2633?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Chengxiang Li updated SPARK-2633: - Attachment: (was: Spark listener enhancement for Hive on Spark job monitor and statistic.docx) > enhance spark listener API to gather more spark job information > --- > > Key: SPARK-2633 > URL: https://issues.apache.org/jira/browse/SPARK-2633 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li >Priority: Critical > Labels: hive > Attachments: Spark listener enhancement for Hive on Spark job monitor > and statistic.docx > > > Based on Hive on Spark job status monitoring and statistic collection > requirement, try to enhance spark listener API to gather more spark job > information. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-2321) Design a proper progress reporting & event listener API
[ https://issues.apache.org/jira/browse/SPARK-2321?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14121085#comment-14121085 ] Chengxiang Li commented on SPARK-2321: -- I collect some hive side requirement here, which should be helpful for spark job status and statistic API design. Hive should be able to get the following job status information through Spark job status API. 1. job identifier 2. current job execution state, should include RUNNING/SUCCEEDED/FAILED/KILLED. 3. running/failed/killed/total task number on job level. 4. stage identifier 5. stage state, should include RUNNING/SUCCEEDED/FAILED/KILLED 6. running/failed/killed/total task number on stage level. MR/Tez use Counter to collect statistic information, similiar to MR/Tez Counter, it would be better if Spark job statistic API organize statistic information with: 1. group same kind statistic information by groupName. 2. displayName for both group and statistic information which would uniform print string for frontend(Web UI/Hive CLI/...). > Design a proper progress reporting & event listener API > --- > > Key: SPARK-2321 > URL: https://issues.apache.org/jira/browse/SPARK-2321 > Project: Spark > Issue Type: Improvement > Components: Java API, Spark Core >Affects Versions: 1.0.0 >Reporter: Reynold Xin >Assignee: Reynold Xin >Priority: Critical > > This is a ticket to track progress on redesigning the SparkListener and > JobProgressListener API. > There are multiple problems with the current design, including: > 0. I'm not sure if the API is usable in Java (there are at least some enums > we used in Scala and a bunch of case classes that might complicate things). > 1. The whole API is marked as DeveloperApi, because we haven't paid a lot of > attention to it yet. Something as important as progress reporting deserves a > more stable API. > 2. There is no easy way to connect jobs with stages. Similarly, there is no > easy way to connect job groups with jobs / stages. > 3. JobProgressListener itself has no encapsulation at all. States can be > arbitrarily mutated by external programs. Variable names are sort of randomly > decided and inconsistent. > We should just revisit these and propose a new, concrete design. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-2321) Design a proper progress reporting & event listener API
[ https://issues.apache.org/jira/browse/SPARK-2321?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14121173#comment-14121173 ] Chengxiang Li commented on SPARK-2321: -- I'm not sure whether i understand you right, here is my thought about the API design: # The JobStatus/JobStatistic API only contains getter method. # JobProgressListener contains variables of JobStatusImpl/JobStatisticImpl. # DagScheduler post events to JobProgressListener through listener bus. # Caller get JobStatusImpl/JobStatisticImpl from JobProgressListener with updated state. So i think it should be a pull style API. > Design a proper progress reporting & event listener API > --- > > Key: SPARK-2321 > URL: https://issues.apache.org/jira/browse/SPARK-2321 > Project: Spark > Issue Type: Improvement > Components: Java API, Spark Core >Affects Versions: 1.0.0 >Reporter: Reynold Xin >Assignee: Reynold Xin >Priority: Critical > > This is a ticket to track progress on redesigning the SparkListener and > JobProgressListener API. > There are multiple problems with the current design, including: > 0. I'm not sure if the API is usable in Java (there are at least some enums > we used in Scala and a bunch of case classes that might complicate things). > 1. The whole API is marked as DeveloperApi, because we haven't paid a lot of > attention to it yet. Something as important as progress reporting deserves a > more stable API. > 2. There is no easy way to connect jobs with stages. Similarly, there is no > easy way to connect job groups with jobs / stages. > 3. JobProgressListener itself has no encapsulation at all. States can be > arbitrarily mutated by external programs. Variable names are sort of randomly > decided and inconsistent. > We should just revisit these and propose a new, concrete design. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-3543) Write TaskContext in Java and expose it through a static accessor
[ https://issues.apache.org/jira/browse/SPARK-3543?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14135016#comment-14135016 ] Chengxiang Li commented on SPARK-3543: -- I think this would solve SPARK-2895 as well, glad to have it. > Write TaskContext in Java and expose it through a static accessor > - > > Key: SPARK-3543 > URL: https://issues.apache.org/jira/browse/SPARK-3543 > Project: Spark > Issue Type: Bug > Components: Spark Core >Reporter: Patrick Wendell >Assignee: Prashant Sharma >Priority: Critical > > Right now we have these xWithContext methods and it's a bit awkward (for > instance, we don't support accessing taskContext from a normal map or filter > operation). I'd propose the following > 1. Re-write TaskContext in Java - it's a simple class. It can still refer to > the scala version of TaskMetrics. > 2. Have a static method `TaskContext.get()` which will return the current > in-scope TaskContext. Under the hood this uses a thread local variable > similar to SparkEnv that the Executor sets. > 3. Deprecate all of the existing xWithContext methods. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-2321) Design a proper progress reporting & event listener API
[ https://issues.apache.org/jira/browse/SPARK-2321?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14143191#comment-14143191 ] Chengxiang Li commented on SPARK-2321: -- I agree that a stable, immutable, and Java-friendly *info classes should be part of this new API design, and between register a new private SparkListener to collect JobInfo and get it from SparkContext, I think there are 2 more issues which should be resolved. # The TaskInfo/StageInfo should be collected by job, while current Stage/Task SparkListener events does not include any job id info, Scheduler have the information to connect TaskInfo/StageInfo with job, maybe we should redesign the SparkListener event API, and add job id info into Stage/Task event in Scheduler before post it to listener bus. # get job id after submit job and get job info by job id. Currently we could only get job id through a very limited way, execute spark async actions with SimpleFutureAction, > Design a proper progress reporting & event listener API > --- > > Key: SPARK-2321 > URL: https://issues.apache.org/jira/browse/SPARK-2321 > Project: Spark > Issue Type: Improvement > Components: Java API, Spark Core >Affects Versions: 1.0.0 >Reporter: Reynold Xin >Assignee: Josh Rosen >Priority: Critical > > This is a ticket to track progress on redesigning the SparkListener and > JobProgressListener API. > There are multiple problems with the current design, including: > 0. I'm not sure if the API is usable in Java (there are at least some enums > we used in Scala and a bunch of case classes that might complicate things). > 1. The whole API is marked as DeveloperApi, because we haven't paid a lot of > attention to it yet. Something as important as progress reporting deserves a > more stable API. > 2. There is no easy way to connect jobs with stages. Similarly, there is no > easy way to connect job groups with jobs / stages. > 3. JobProgressListener itself has no encapsulation at all. States can be > arbitrarily mutated by external programs. Variable names are sort of randomly > decided and inconsistent. > We should just revisit these and propose a new, concrete design. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Created] (SPARK-1442) Add Window function support
Chengxiang Li created SPARK-1442: Summary: Add Window function support Key: SPARK-1442 URL: https://issues.apache.org/jira/browse/SPARK-1442 Project: Spark Issue Type: New Feature Components: SQL Reporter: Chengxiang Li similiar to Hive, add window function support for catalyst. https://issues.apache.org/jira/browse/HIVE-4197 https://issues.apache.org/jira/browse/HIVE-896 -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Created] (SPARK-2633) support register spark listener to listener bus with Java API
Chengxiang Li created SPARK-2633: Summary: support register spark listener to listener bus with Java API Key: SPARK-2633 URL: https://issues.apache.org/jira/browse/SPARK-2633 Project: Spark Issue Type: New Feature Components: Java API Reporter: Chengxiang Li Currently user can only register spark listener with Scala API, we should add this feature to Java API as well. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (SPARK-2633) support register spark listener to listener bus with Java API
[ https://issues.apache.org/jira/browse/SPARK-2633?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14071310#comment-14071310 ] Chengxiang Li commented on SPARK-2633: -- cc [~rxin] [~xuefuz] > support register spark listener to listener bus with Java API > - > > Key: SPARK-2633 > URL: https://issues.apache.org/jira/browse/SPARK-2633 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li > > Currently user can only register spark listener with Scala API, we should add > this feature to Java API as well. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Created] (SPARK-2636) no where to get job identifier while submit spark job through spark API
Chengxiang Li created SPARK-2636: Summary: no where to get job identifier while submit spark job through spark API Key: SPARK-2636 URL: https://issues.apache.org/jira/browse/SPARK-2636 Project: Spark Issue Type: New Feature Reporter: Chengxiang Li In Hive on Spark, we want to track spark job status through Spark API, the basic idea is as following: # create an hive-specified spark listener and register it to spark listener bus. # hive-specified spark listener generate job status by spark listener events. # hive driver track job status through hive-specified spark listener. the current problem is that hive driver need job identifier to track specified job status through spark listener, but there is no spark API to get job identifier(like job id) while submit spark job. I think other project whoever try to track job status with spark API would suffer from this as well. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (SPARK-2636) no where to get job identifier while submit spark job through spark API
[ https://issues.apache.org/jira/browse/SPARK-2636?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Chengxiang Li updated SPARK-2636: - Description: In Hive on Spark, we want to track spark job status through Spark API, the basic idea is as following: # create an hive-specified spark listener and register it to spark listener bus. # hive-specified spark listener generate job status by spark listener events. # hive driver track job status through hive-specified spark listener. the current problem is that hive driver need job identifier to track specified job status through spark listener, but there is no spark API to get job identifier(like job id) while submit spark job. I think other project whoever try to track job status with spark API would suffer from this as well. was: In Hive on Spark, we want to track spark job status through Spark API, the basic idea is as following: # create an hive-specified spark listener and register it to spark listener bus. # hive-specified spark listener generate job status by spark listener events. # hive driver track job status through hive-specified spark listener. the current problem is that hive driver need job identifier to track specified job status through spark listener, but there is no spark API to get job identifier(like job id) while submit spark job. I think other project whoever try to track job status with spark API would suffer from this as well. > no where to get job identifier while submit spark job through spark API > --- > > Key: SPARK-2636 > URL: https://issues.apache.org/jira/browse/SPARK-2636 > Project: Spark > Issue Type: New Feature >Reporter: Chengxiang Li > > In Hive on Spark, we want to track spark job status through Spark API, the > basic idea is as following: > # create an hive-specified spark listener and register it to spark listener > bus. > # hive-specified spark listener generate job status by spark listener events. > # hive driver track job status through hive-specified spark listener. > the current problem is that hive driver need job identifier to track > specified job status through spark listener, but there is no spark API to get > job identifier(like job id) while submit spark job. > I think other project whoever try to track job status with spark API would > suffer from this as well. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (SPARK-2636) no where to get job identifier while submit spark job through spark API
[ https://issues.apache.org/jira/browse/SPARK-2636?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14071311#comment-14071311 ] Chengxiang Li commented on SPARK-2636: -- cc [~rxin] [~xuefuz] > no where to get job identifier while submit spark job through spark API > --- > > Key: SPARK-2636 > URL: https://issues.apache.org/jira/browse/SPARK-2636 > Project: Spark > Issue Type: New Feature >Reporter: Chengxiang Li > > In Hive on Spark, we want to track spark job status through Spark API, the > basic idea is as following: > # create an hive-specified spark listener and register it to spark listener > bus. > # hive-specified spark listener generate job status by spark listener events. > # hive driver track job status through hive-specified spark listener. > the current problem is that hive driver need job identifier to track > specified job status through spark listener, but there is no spark API to get > job identifier(like job id) while submit spark job. > I think other project whoever try to track job status with spark API would > suffer from this as well. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (SPARK-2636) no where to get job identifier while submit spark job through spark API
[ https://issues.apache.org/jira/browse/SPARK-2636?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Chengxiang Li updated SPARK-2636: - Description: In Hive on Spark, we want to track spark job status through Spark API, the basic idea is as following: # create an hive-specified spark listener and register it to spark listener bus. # hive-specified spark listener generate job status by spark listener events. # hive driver track job status through hive-specified spark listener. the current problem is that hive driver need job identifier to track specified job status through spark listener, but there is no spark API to get job identifier(like job id) while submit spark job. I think other project whoever try to track job status with spark API would suffer from this as well. was: In Hive on Spark, we want to track spark job status through Spark API, the basic idea is as following: # create an hive-specified spark listener and register it to spark listener bus. # hive-specified spark listener generate job status by spark listener events. # hive driver track job status through hive-specified spark listener. the current problem is that hive driver need job identifier to track specified job status through spark listener, but there is no spark API to get job identifier(like job id) while submit spark job. I think other project whoever try to track job status with spark API would suffer from this as well. > no where to get job identifier while submit spark job through spark API > --- > > Key: SPARK-2636 > URL: https://issues.apache.org/jira/browse/SPARK-2636 > Project: Spark > Issue Type: New Feature >Reporter: Chengxiang Li > > In Hive on Spark, we want to track spark job status through Spark API, the > basic idea is as following: > # create an hive-specified spark listener and register it to spark listener > bus. > # hive-specified spark listener generate job status by spark listener events. > # hive driver track job status through hive-specified spark listener. > the current problem is that hive driver need job identifier to track > specified job status through spark listener, but there is no spark API to get > job identifier(like job id) while submit spark job. > I think other project whoever try to track job status with spark API would > suffer from this as well. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (SPARK-2633) support register spark listener to listener bus with Java API
[ https://issues.apache.org/jira/browse/SPARK-2633?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14071402#comment-14071402 ] Chengxiang Li commented on SPARK-2633: -- For Hive job status monitor, spark listener monitor less job status than expected: current * monitor events: onJobStart, onJobEnd(JobSucceeded/JobFailed) * available status: Started/Succeeded/Failed expect: * monitor events: +onJobSubmitted, onJobStart, onJobEnd(JobSucceeded/JobFailed/+JobKilled) * available status: Submitted/Started/Succeeded/Failed/Killed > support register spark listener to listener bus with Java API > - > > Key: SPARK-2633 > URL: https://issues.apache.org/jira/browse/SPARK-2633 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li > > Currently user can only register spark listener with Scala API, we should add > this feature to Java API as well. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (SPARK-2633) support register spark listener to listener bus with Java API
[ https://issues.apache.org/jira/browse/SPARK-2633?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14071478#comment-14071478 ] Chengxiang Li commented on SPARK-2633: -- add 2 more: # StageInfo class is not well designed to get stage state, should contain something like StageEndReason. # it would be lovely if spark java API could transfer scala collection into java collection, not by user. for example, stageIds: Seq\[Int] in SparkListenerJobStart. > support register spark listener to listener bus with Java API > - > > Key: SPARK-2633 > URL: https://issues.apache.org/jira/browse/SPARK-2633 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li > > Currently user can only register spark listener with Scala API, we should add > this feature to Java API as well. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (SPARK-2633) support register spark listener to listener bus with Java API
[ https://issues.apache.org/jira/browse/SPARK-2633?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14072807#comment-14072807 ] Chengxiang Li commented on SPARK-2633: -- Thanks [~vanzin], that's a very good point. Stand at the side of Hive job monitor, i think we may need 2 kinds of support from spark: # Full feature support of job monitor by spark listener, as i mentioned in previous comments, we are unable to fetch enough messages by current spark listener yet. # Java friendly API like what spark has done with RDD, instead of just adding a proxy method to add spark listener in JavaSparkContext. #1 is needed for hive on spark job monitor feature. #2 is not neccessarily required as scala style spark listener is workable through not graceful in java, but it would be definite expected if spark can supply a java native spark listener API. > support register spark listener to listener bus with Java API > - > > Key: SPARK-2633 > URL: https://issues.apache.org/jira/browse/SPARK-2633 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li > > Currently user can only register spark listener with Scala API, we should add > this feature to Java API as well. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (SPARK-2633) support register spark listener to listener bus with Java API
[ https://issues.apache.org/jira/browse/SPARK-2633?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14080425#comment-14080425 ] Chengxiang Li commented on SPARK-2633: -- Thanks, [~vanzin] {quote} Registering listeners through Java API: although not super-clean, it can be done today easily. SparkListener is a pretty clean interface that should be easily implementable in Java-land. You can register it using JavaSparkContext.sc().addSparkListener(). Is that enough? {quote} it's workable, but maybe add a proxy method such as JavaSparkContext.addSparkListener() is more meaningful for java spark listener API, what do you think? {quote} Events: SparkListenerJobStart is actually issued at job submission time (see DAGScheduler.scala::handleJobSubmitted, DAGScheduler.scala::submitJob. Not sure there's any other good substitute for it; the job has technically started - the DAGScheduler is processing it, and at some point might start executing tasks. I guess it depends on what is it that you're trying to achieve here... {quote} Previously i assumed that spark job has submitted state before running, while it seems does not. DAGScheduler would break up spark job into tasks after submit job and add to pending task list, waiting executor offer resource to execute these tasks. So spark job wait for running in task level instead of job level. For “TaskStarted” and "TaskSubmitted", do you think it's reasonable to use "TaskSubmitted" to represent the task state between task get submitted(TaskSchedulerImpl.scala::submitTasks()) and task get resource allocated(TaskSchedulerImpl.scala::resourceOffers())? Besides, current StageInfo is hold too little information, we may need to add state object inside like StageEndReason(success/failed/killed/resubmitted) to represent different kinds of stage end info.. > support register spark listener to listener bus with Java API > - > > Key: SPARK-2633 > URL: https://issues.apache.org/jira/browse/SPARK-2633 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li > > Currently user can only register spark listener with Scala API, we should add > this feature to Java API as well. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (SPARK-2633) support register spark listener to listener bus with Java API
[ https://issues.apache.org/jira/browse/SPARK-2633?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Chengxiang Li updated SPARK-2633: - Attachment: Spark listener enhancement for Hive on Spark job monitor and statistic.docx I add a doc to collect requirement from hive on spark side, it may looks mussy in the comments. we could keep on discussing based on this file. > support register spark listener to listener bus with Java API > - > > Key: SPARK-2633 > URL: https://issues.apache.org/jira/browse/SPARK-2633 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li > Attachments: Spark listener enhancement for Hive on Spark job monitor > and statistic.docx > > > Currently user can only register spark listener with Scala API, we should add > this feature to Java API as well. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (SPARK-2636) no where to get job identifier while submit spark job through spark API
[ https://issues.apache.org/jira/browse/SPARK-2636?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14085674#comment-14085674 ] Chengxiang Li commented on SPARK-2636: -- Thanks, [~vanzin]. Hive on Spark mainly try to track job level status which could be collected by job/stage/task events, so perhaps environment level events does not help a lot here. Currently our problem is that there is no Spark API to get job identifier, Spark generate jobId internally while submit job, it just not exposed in API level. How would you expose job id in the PR? > no where to get job identifier while submit spark job through spark API > --- > > Key: SPARK-2636 > URL: https://issues.apache.org/jira/browse/SPARK-2636 > Project: Spark > Issue Type: New Feature >Reporter: Chengxiang Li > > In Hive on Spark, we want to track spark job status through Spark API, the > basic idea is as following: > # create an hive-specified spark listener and register it to spark listener > bus. > # hive-specified spark listener generate job status by spark listener events. > # hive driver track job status through hive-specified spark listener. > the current problem is that hive driver need job identifier to track > specified job status through spark listener, but there is no spark API to get > job identifier(like job id) while submit spark job. > I think other project whoever try to track job status with spark API would > suffer from this as well. -- This message was sent by Atlassian JIRA (v6.2#6252) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-2636) no where to get job identifier while submit spark job through spark API
[ https://issues.apache.org/jira/browse/SPARK-2636?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14087102#comment-14087102 ] Chengxiang Li commented on SPARK-2636: -- {quote} There are two ways I think. One is for DAGScheduler.runJob to return an integer (or long) id for the job. An alternative, which I think is better and relates to SPARK-2321, is for runJob to return some Job object that has information about the id and can be queried about progress. {quote} DAGScheduler is Spark internal class, User can hardly use it directly. I like your second idea, return a Job info object while submit spark job in SparkContext(JavaSparkContext in this case) or RDD level. Actually AsyncRDDActions has done part of this work, I think it maybe a good place to fix this issue. > no where to get job identifier while submit spark job through spark API > --- > > Key: SPARK-2636 > URL: https://issues.apache.org/jira/browse/SPARK-2636 > Project: Spark > Issue Type: New Feature >Reporter: Chengxiang Li > > In Hive on Spark, we want to track spark job status through Spark API, the > basic idea is as following: > # create an hive-specified spark listener and register it to spark listener > bus. > # hive-specified spark listener generate job status by spark listener events. > # hive driver track job status through hive-specified spark listener. > the current problem is that hive driver need job identifier to track > specified job status through spark listener, but there is no spark API to get > job identifier(like job id) while submit spark job. > I think other project whoever try to track job status with spark API would > suffer from this as well. -- This message was sent by Atlassian JIRA (v6.2#6252) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Created] (SPARK-2895) Support mapPartitionsWithContext in Spark Java API
Chengxiang Li created SPARK-2895: Summary: Support mapPartitionsWithContext in Spark Java API Key: SPARK-2895 URL: https://issues.apache.org/jira/browse/SPARK-2895 Project: Spark Issue Type: New Feature Components: Java API Reporter: Chengxiang Li This is a requirement from Hive on Spark, mapPartitionsWithContext only exists in Spark Scala API, we expect to access from Spark Java API. -- This message was sent by Atlassian JIRA (v6.2#6252) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-2895) Support mapPartitionsWithContext in Spark Java API
[ https://issues.apache.org/jira/browse/SPARK-2895?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14088736#comment-14088736 ] Chengxiang Li commented on SPARK-2895: -- cc [~rxin] [~brocknoland] [~szehon] > Support mapPartitionsWithContext in Spark Java API > -- > > Key: SPARK-2895 > URL: https://issues.apache.org/jira/browse/SPARK-2895 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li > > This is a requirement from Hive on Spark, mapPartitionsWithContext only > exists in Spark Scala API, we expect to access from Spark Java API. -- This message was sent by Atlassian JIRA (v6.2#6252) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-2895) Support mapPartitionsWithContext in Spark Java API
[ https://issues.apache.org/jira/browse/SPARK-2895?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14088895#comment-14088895 ] Chengxiang Li commented on SPARK-2895: -- {quote} Can we add the label "hive" to all the tickets related to the Hive on Spark project in the future? This should be fairly easy to do. {quote} Thanks, [~rxin], that make sense, i would label all hive on spark related issues. > Support mapPartitionsWithContext in Spark Java API > -- > > Key: SPARK-2895 > URL: https://issues.apache.org/jira/browse/SPARK-2895 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li > Labels: hive > > This is a requirement from Hive on Spark, mapPartitionsWithContext only > exists in Spark Scala API, we expect to access from Spark Java API. -- This message was sent by Atlassian JIRA (v6.2#6252) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-2633) support register spark listener to listener bus with Java API
[ https://issues.apache.org/jira/browse/SPARK-2633?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Chengxiang Li updated SPARK-2633: - Labels: hive (was: ) > support register spark listener to listener bus with Java API > - > > Key: SPARK-2633 > URL: https://issues.apache.org/jira/browse/SPARK-2633 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li >Priority: Critical > Labels: hive > Attachments: Spark listener enhancement for Hive on Spark job monitor > and statistic.docx > > > Currently user can only register spark listener with Scala API, we should add > this feature to Java API as well. -- This message was sent by Atlassian JIRA (v6.2#6252) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-2636) no where to get job identifier while submit spark job through spark API
[ https://issues.apache.org/jira/browse/SPARK-2636?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Chengxiang Li updated SPARK-2636: - Component/s: Java API > no where to get job identifier while submit spark job through spark API > --- > > Key: SPARK-2636 > URL: https://issues.apache.org/jira/browse/SPARK-2636 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li > Labels: hive > > In Hive on Spark, we want to track spark job status through Spark API, the > basic idea is as following: > # create an hive-specified spark listener and register it to spark listener > bus. > # hive-specified spark listener generate job status by spark listener events. > # hive driver track job status through hive-specified spark listener. > the current problem is that hive driver need job identifier to track > specified job status through spark listener, but there is no spark API to get > job identifier(like job id) while submit spark job. > I think other project whoever try to track job status with spark API would > suffer from this as well. -- This message was sent by Atlassian JIRA (v6.2#6252) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Updated] (SPARK-2636) no where to get job identifier while submit spark job through spark API
[ https://issues.apache.org/jira/browse/SPARK-2636?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Chengxiang Li updated SPARK-2636: - Labels: hive (was: ) > no where to get job identifier while submit spark job through spark API > --- > > Key: SPARK-2636 > URL: https://issues.apache.org/jira/browse/SPARK-2636 > Project: Spark > Issue Type: New Feature > Components: Java API >Reporter: Chengxiang Li > Labels: hive > > In Hive on Spark, we want to track spark job status through Spark API, the > basic idea is as following: > # create an hive-specified spark listener and register it to spark listener > bus. > # hive-specified spark listener generate job status by spark listener events. > # hive driver track job status through hive-specified spark listener. > the current problem is that hive driver need job identifier to track > specified job status through spark listener, but there is no spark API to get > job identifier(like job id) while submit spark job. > I think other project whoever try to track job status with spark API would > suffer from this as well. -- This message was sent by Atlassian JIRA (v6.2#6252) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Created] (SPARK-5377) Dynamically add jar into Spark Driver's classpath.
Chengxiang Li created SPARK-5377: Summary: Dynamically add jar into Spark Driver's classpath. Key: SPARK-5377 URL: https://issues.apache.org/jira/browse/SPARK-5377 Project: Spark Issue Type: Improvement Affects Versions: 1.2.0 Reporter: Chengxiang Li Spark support dynamically add jar to executor classpath through SparkContext::addJar(), while it does not support dynamically add jar into driver classpath. In most case(if not all the case), user dynamically add jar with SparkContext::addJar() because some classes from the jar would be referred in upcoming Spark job, which means the classes need to be loaded in Spark driver side either,e.g during serialization. I think it make sense to add an API to add jar into driver classpath, or just make it available in SparkContext::addJar(). HIVE-9410 is a real case from Hive on Spark. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-5377) Dynamically add jar into Spark Driver's classpath.
[ https://issues.apache.org/jira/browse/SPARK-5377?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14288783#comment-14288783 ] Chengxiang Li commented on SPARK-5377: -- cc [~xuefuz], [~rxin], [~Grace Huang]. > Dynamically add jar into Spark Driver's classpath. > -- > > Key: SPARK-5377 > URL: https://issues.apache.org/jira/browse/SPARK-5377 > Project: Spark > Issue Type: Improvement >Affects Versions: 1.2.0 >Reporter: Chengxiang Li > > Spark support dynamically add jar to executor classpath through > SparkContext::addJar(), while it does not support dynamically add jar into > driver classpath. In most case(if not all the case), user dynamically add jar > with SparkContext::addJar() because some classes from the jar would be > referred in upcoming Spark job, which means the classes need to be loaded in > Spark driver side either,e.g during serialization. I think it make sense to > add an API to add jar into driver classpath, or just make it available in > SparkContext::addJar(). HIVE-9410 is a real case from Hive on Spark. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-5439) Expose yarn app id for yarn mode
[ https://issues.apache.org/jira/browse/SPARK-5439?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14294872#comment-14294872 ] Chengxiang Li commented on SPARK-5439: -- I think the gap here is that, when launch a Hive CLI with Spark deployed on YARN, from the YARN webUI page, the user want to know which YARN application is allocated for his hive session, then he can check the status of his application through ApplicationMaster tracking UI. [~vanzin], is the YARN application ID same as SparkContext applicationID? From my experience, seems not. The YARN web UI shows that the YARN application ID schema is like:application_\*, and SparkContext applicationID is set at SchedulerBackend#26 with schema spark-application-*. > Expose yarn app id for yarn mode > > > Key: SPARK-5439 > URL: https://issues.apache.org/jira/browse/SPARK-5439 > Project: Spark > Issue Type: Improvement > Components: YARN >Affects Versions: 1.0.0 >Reporter: bc Wong > > When submitting Spark apps on YARN, the caller should be able to get back the > YARN app ID programmatically. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org