[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user AmplabJenkins commented on the pull request: https://github.com/apache/spark/pull/219#issuecomment-38534171 One or more automated tests failed Refer to this link for build results: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/13423/ --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user kayousterhout commented on the pull request: https://github.com/apache/spark/pull/219#issuecomment-38592694 LGTM --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user AmplabJenkins commented on the pull request: https://github.com/apache/spark/pull/219#issuecomment-38600307 One or more automated tests failed Refer to this link for build results: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/13437/ --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user AmplabJenkins commented on the pull request: https://github.com/apache/spark/pull/219#issuecomment-38600305 Merged build finished. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user shivaram commented on the pull request: https://github.com/apache/spark/pull/219#issuecomment-38601404 Jenkins, retest this please --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user shivaram commented on the pull request: https://github.com/apache/spark/pull/219#issuecomment-38601384 Failure was in recovery with file input stream.recovery with file input stream -- Something that I think is completely unrelated to this change. So lets try again ... --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user AmplabJenkins commented on the pull request: https://github.com/apache/spark/pull/219#issuecomment-38601510 Merged build triggered. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user AmplabJenkins commented on the pull request: https://github.com/apache/spark/pull/219#issuecomment-38608422 All automated tests passed. Refer to this link for build results: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/13440/ --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user mridulm commented on the pull request: https://github.com/apache/spark/pull/219#issuecomment-38608715 Looks good, thanks for the change - makes things cleaner. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
GitHub user shivaram opened a pull request: https://github.com/apache/spark/pull/219 Fix scheduler to account for tasks using 1 CPUs. Move CPUS_PER_TASK to TaskSchedulerImpl as the value is a constant and use it in both Mesos and CoarseGrained scheduler backends. Thanks @kayousterhout for the design discussion You can merge this pull request into a Git repository by running: $ git pull https://github.com/shivaram/spark-1 multi-cpus Alternatively you can review and apply these changes as the patch at: https://github.com/apache/spark/pull/219.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #219 commit 647bc4582bb350faf69d1969072a2cc79f2f553e Author: Shivaram Venkataraman shiva...@eecs.berkeley.edu Date: 2014-03-25T01:53:03Z Fix scheduler to account for tasks using 1 CPUs. Move CPUS_PER_TASK to TaskSchedulerImpl as the value is a constant and use it in both Mesos and CoarseGrained scheduler backends. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user pwendell commented on a diff in the pull request: https://github.com/apache/spark/pull/219#discussion_r10915292 --- Diff: core/src/main/scala/org/apache/spark/scheduler/TaskSchedulerImpl.scala --- @@ -62,6 +62,9 @@ private[spark] class TaskSchedulerImpl( // Threshold above which we warn user initial TaskSet may be starved val STARVATION_TIMEOUT = conf.getLong(spark.starvation.timeout, 15000) + // CPUs to request per task + val CPUS_PER_TASK = conf.getInt(spark.task.cpus, 1) --- End diff -- Is this intentionally left as an undocumented parameter? --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user shivaram commented on a diff in the pull request: https://github.com/apache/spark/pull/219#discussion_r10915339 --- Diff: core/src/main/scala/org/apache/spark/scheduler/TaskSchedulerImpl.scala --- @@ -62,6 +62,9 @@ private[spark] class TaskSchedulerImpl( // Threshold above which we warn user initial TaskSet may be starved val STARVATION_TIMEOUT = conf.getLong(spark.starvation.timeout, 15000) + // CPUs to request per task + val CPUS_PER_TASK = conf.getInt(spark.task.cpus, 1) --- End diff -- Hmm - we could document this in docs/configuration.md -- but I don't think this is a commonly used flag, and it has been around for a while I guess. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user shivaram commented on a diff in the pull request: https://github.com/apache/spark/pull/219#discussion_r10915407 --- Diff: core/src/main/scala/org/apache/spark/scheduler/TaskSchedulerImpl.scala --- @@ -62,6 +62,9 @@ private[spark] class TaskSchedulerImpl( // Threshold above which we warn user initial TaskSet may be starved val STARVATION_TIMEOUT = conf.getLong(spark.starvation.timeout, 15000) + // CPUs to request per task + val CPUS_PER_TASK = conf.getInt(spark.task.cpus, 1) --- End diff -- I guess documenting it is better -- added a commit for that. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user shivaram commented on the pull request: https://github.com/apache/spark/pull/219#issuecomment-38524982 FYI - there was a test failure in TaskSetManagerSuite as a unit test was checking for availableCpus being zero. So I added back a non-zero check in TaskSetManager. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user AmplabJenkins commented on the pull request: https://github.com/apache/spark/pull/219#issuecomment-38526638 One or more automated tests failed Refer to this link for build results: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/13414/ --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user AmplabJenkins commented on the pull request: https://github.com/apache/spark/pull/219#issuecomment-38526636 Merged build finished. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user AmplabJenkins commented on the pull request: https://github.com/apache/spark/pull/219#issuecomment-38526736 Merged build triggered. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user AmplabJenkins commented on the pull request: https://github.com/apache/spark/pull/219#issuecomment-38527171 One or more automated tests failed Refer to this link for build results: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/13417/ --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user AmplabJenkins commented on the pull request: https://github.com/apache/spark/pull/219#issuecomment-38527170 Merged build finished. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user shivaram commented on the pull request: https://github.com/apache/spark/pull/219#issuecomment-38529622 Jenkins, retest this please --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user AmplabJenkins commented on the pull request: https://github.com/apache/spark/pull/219#issuecomment-38529627 Merged build started. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user AmplabJenkins commented on the pull request: https://github.com/apache/spark/pull/219#issuecomment-38529626 Merged build triggered. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user kayousterhout commented on a diff in the pull request: https://github.com/apache/spark/pull/219#discussion_r10917822 --- Diff: core/src/main/scala/org/apache/spark/scheduler/TaskSetManager.scala --- @@ -388,7 +385,7 @@ private[spark] class TaskSetManager( maxLocality: TaskLocality.TaskLocality) : Option[TaskDescription] = { -if (!isZombie availableCpus = CPUS_PER_TASK) { +if (!isZombie availableCpus 0) { --- End diff -- Can you just remove the availableCpus parameter? It doesn't look like it's used anymore --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user AmplabJenkins commented on the pull request: https://github.com/apache/spark/pull/219#issuecomment-38531549 All automated tests passed. Refer to this link for build results: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/13421/ --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user shivaram commented on a diff in the pull request: https://github.com/apache/spark/pull/219#discussion_r10918164 --- Diff: core/src/main/scala/org/apache/spark/scheduler/TaskSetManager.scala --- @@ -388,7 +385,7 @@ private[spark] class TaskSetManager( maxLocality: TaskLocality.TaskLocality) : Option[TaskDescription] = { -if (!isZombie availableCpus = CPUS_PER_TASK) { +if (!isZombie availableCpus 0) { --- End diff -- Makes sense -- Removed it now --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] spark pull request: Fix scheduler to account for tasks using 1 C...
Github user pwendell commented on a diff in the pull request: https://github.com/apache/spark/pull/219#discussion_r10918370 --- Diff: core/src/main/scala/org/apache/spark/scheduler/TaskSchedulerImpl.scala --- @@ -62,6 +62,9 @@ private[spark] class TaskSchedulerImpl( // Threshold above which we warn user initial TaskSet may be starved val STARVATION_TIMEOUT = conf.getLong(spark.starvation.timeout, 15000) + // CPUs to request per task + val CPUS_PER_TASK = conf.getInt(spark.task.cpus, 1) --- End diff -- Ah I see that it wasn't part of your change. Anyways it might make sense to document it. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---