[jira] [Commented] (SPARK-27499) Support mapping spark.local.dir to hostPath volume
[ https://issues.apache.org/jira/browse/SPARK-27499?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16885666#comment-16885666 ] Junjie Chen commented on SPARK-27499: - Hi [~vanzin] There is an opened Jira SPARK-28042 for this. > Support mapping spark.local.dir to hostPath volume > -- > > Key: SPARK-27499 > URL: https://issues.apache.org/jira/browse/SPARK-27499 > Project: Spark > Issue Type: Improvement > Components: Kubernetes >Affects Versions: 3.0.0 >Reporter: Junjie Chen >Priority: Minor > > Currently, the k8s executor builder mount spark.local.dir as emptyDir or > memory, it should satisfy some small workload, while in some heavily workload > like TPCDS, both of them can have some problem, such as pods are evicted due > to disk pressure when using emptyDir, and OOM when using tmpfs. > In particular on cloud environment, users may allocate cluster with minimum > configuration and add cloud storage when running workload. In this case, we > can specify multiple elastic storage as spark.local.dir to accelerate the > spilling. -- This message was sent by Atlassian JIRA (v7.6.14#76016) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-27499) Support mapping spark.local.dir to hostPath volume
[ https://issues.apache.org/jira/browse/SPARK-27499?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16885578#comment-16885578 ] Marcelo Vanzin commented on SPARK-27499: I can't see an option to reopen this, so I'll clone it instead. This seems like a simple fix that can at least help people experiment with different storage. > Support mapping spark.local.dir to hostPath volume > -- > > Key: SPARK-27499 > URL: https://issues.apache.org/jira/browse/SPARK-27499 > Project: Spark > Issue Type: Improvement > Components: Kubernetes >Affects Versions: 3.0.0 >Reporter: Junjie Chen >Priority: Minor > > Currently, the k8s executor builder mount spark.local.dir as emptyDir or > memory, it should satisfy some small workload, while in some heavily workload > like TPCDS, both of them can have some problem, such as pods are evicted due > to disk pressure when using emptyDir, and OOM when using tmpfs. > In particular on cloud environment, users may allocate cluster with minimum > configuration and add cloud storage when running workload. In this case, we > can specify multiple elastic storage as spark.local.dir to accelerate the > spilling. -- This message was sent by Atlassian JIRA (v7.6.14#76016) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-27499) Support mapping spark.local.dir to hostPath volume
[ https://issues.apache.org/jira/browse/SPARK-27499?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16861811#comment-16861811 ] Dongjoon Hyun commented on SPARK-27499: --- Got it. Sorry for closing this issue, [~junjie]. I'll take a look a little bit more and reopen (or clone) this issue with the same reporter name. > Support mapping spark.local.dir to hostPath volume > -- > > Key: SPARK-27499 > URL: https://issues.apache.org/jira/browse/SPARK-27499 > Project: Spark > Issue Type: Improvement > Components: Kubernetes >Affects Versions: 3.0.0 >Reporter: Junjie Chen >Priority: Minor > > Currently, the k8s executor builder mount spark.local.dir as emptyDir or > memory, it should satisfy some small workload, while in some heavily workload > like TPCDS, both of them can have some problem, such as pods are evicted due > to disk pressure when using emptyDir, and OOM when using tmpfs. > In particular on cloud environment, users may allocate cluster with minimum > configuration and add cloud storage when running workload. In this case, we > can specify multiple elastic storage as spark.local.dir to accelerate the > spilling. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-27499) Support mapping spark.local.dir to hostPath volume
[ https://issues.apache.org/jira/browse/SPARK-27499?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16861707#comment-16861707 ] Junjie Chen commented on SPARK-27499: - Yes, In KubernetesExecutorBuilder.scala, the LocalDrisFeatureStep is built before MountVolumesFeatureStep which means we cannot use any volumes mount later. I think we should build localDirsFeature at last, so that we can check if directories in SPARK_LOCAL_DIRS are set to volumes mounted either hostPath, PV, or others may support later and use that as local storage. With that way, we can utilize specified media to improve the local storage performance instead of just emptyDir which is just a ephemeral directory on node. > Support mapping spark.local.dir to hostPath volume > -- > > Key: SPARK-27499 > URL: https://issues.apache.org/jira/browse/SPARK-27499 > Project: Spark > Issue Type: Improvement > Components: Kubernetes >Affects Versions: 3.0.0 >Reporter: Junjie Chen >Priority: Minor > > Currently, the k8s executor builder mount spark.local.dir as emptyDir or > memory, it should satisfy some small workload, while in some heavily workload > like TPCDS, both of them can have some problem, such as pods are evicted due > to disk pressure when using emptyDir, and OOM when using tmpfs. > In particular on cloud environment, users may allocate cluster with minimum > configuration and add cloud storage when running workload. In this case, we > can specify multiple elastic storage as spark.local.dir to accelerate the > spilling. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-27499) Support mapping spark.local.dir to hostPath volume
[ https://issues.apache.org/jira/browse/SPARK-27499?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16860648#comment-16860648 ] Dongjoon Hyun commented on SPARK-27499: --- Oh, do you mean `SPARK_LOCAL_DIRS` doesn't work on the PV, [~junjie]? > Support mapping spark.local.dir to hostPath volume > -- > > Key: SPARK-27499 > URL: https://issues.apache.org/jira/browse/SPARK-27499 > Project: Spark > Issue Type: Improvement > Components: Kubernetes >Affects Versions: 3.0.0 >Reporter: Junjie Chen >Priority: Minor > > Currently, the k8s executor builder mount spark.local.dir as emptyDir or > memory, it should satisfy some small workload, while in some heavily workload > like TPCDS, both of them can have some problem, such as pods are evicted due > to disk pressure when using emptyDir, and OOM when using tmpfs. > In particular on cloud environment, users may allocate cluster with minimum > configuration and add cloud storage when running workload. In this case, we > can specify multiple elastic storage as spark.local.dir to accelerate the > spilling. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-27499) Support mapping spark.local.dir to hostPath volume
[ https://issues.apache.org/jira/browse/SPARK-27499?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16860598#comment-16860598 ] Dongjoon Hyun commented on SPARK-27499: --- No, I think you didn't read the patch of that, [~junjie]. Could you read the patch? You can start from the following line. - https://github.com/apache/spark/pull/21238/files#diff-529fc5c06b9731c1fbda6f3db60b16aaR458 > Support mapping spark.local.dir to hostPath volume > -- > > Key: SPARK-27499 > URL: https://issues.apache.org/jira/browse/SPARK-27499 > Project: Spark > Issue Type: Improvement > Components: Kubernetes >Affects Versions: 3.0.0 >Reporter: Junjie Chen >Priority: Minor > Fix For: 2.4.0 > > > Currently, the k8s executor builder mount spark.local.dir as emptyDir or > memory, it should satisfy some small workload, while in some heavily workload > like TPCDS, both of them can have some problem, such as pods are evicted due > to disk pressure when using emptyDir, and OOM when using tmpfs. > In particular on cloud environment, users may allocate cluster with minimum > configuration and add cloud storage when running workload. In this case, we > can specify multiple elastic storage as spark.local.dir to accelerate the > spilling. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-27499) Support mapping spark.local.dir to hostPath volume
[ https://issues.apache.org/jira/browse/SPARK-27499?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16860519#comment-16860519 ] Junjie Chen commented on SPARK-27499: - Hi, [~dongjoon], I know SPARK_LOCAL_DIRS can be mounted as emptyDir. However, emptyDir just one directory on node. I opened this Jira to track a feature to setting multiple directories to full utilize the nodes' disks bandwidth for spilling, which I think currently it can not be achieve through setting spark.local.dir. Even I set to multiple dirs, they still map to one directory on node. This Jira is intended to use hostPath volumes mounts as spark.local.dir, for exmaple: spark.kubernetes.executor.volumes.hostPath.spark-local-dir-1.mount.path=/data/mnt-x > Support mapping spark.local.dir to hostPath volume > -- > > Key: SPARK-27499 > URL: https://issues.apache.org/jira/browse/SPARK-27499 > Project: Spark > Issue Type: Improvement > Components: Kubernetes >Affects Versions: 3.0.0 >Reporter: Junjie Chen >Priority: Minor > Fix For: 2.4.0 > > > Currently, the k8s executor builder mount spark.local.dir as emptyDir or > memory, it should satisfy some small workload, while in some heavily workload > like TPCDS, both of them can have some problem, such as pods are evicted due > to disk pressure when using emptyDir, and OOM when using tmpfs. > In particular on cloud environment, users may allocate cluster with minimum > configuration and add cloud storage when running workload. In this case, we > can specify multiple elastic storage as spark.local.dir to accelerate the > spilling. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org