[jira] [Commented] (SPARK-28042) Support mapping spark.local.dir to hostPath volume
[ https://issues.apache.org/jira/browse/SPARK-28042?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16925349#comment-16925349 ] Jiaxin Shan commented on SPARK-28042: - [~dongjoon] Thanks! I will try to cherry-pick changes and build a customized version for now. > Support mapping spark.local.dir to hostPath volume > -- > > Key: SPARK-28042 > URL: https://issues.apache.org/jira/browse/SPARK-28042 > Project: Spark > Issue Type: Improvement > Components: Kubernetes >Affects Versions: 3.0.0 >Reporter: Junjie Chen >Assignee: Junjie Chen >Priority: Minor > Fix For: 3.0.0 > > > 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 (v8.3.2#803003) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-28042) Support mapping spark.local.dir to hostPath volume
[ https://issues.apache.org/jira/browse/SPARK-28042?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16925274#comment-16925274 ] Dongjoon Hyun commented on SPARK-28042: --- No, we usually backport bug fixes only. > Support mapping spark.local.dir to hostPath volume > -- > > Key: SPARK-28042 > URL: https://issues.apache.org/jira/browse/SPARK-28042 > Project: Spark > Issue Type: Improvement > Components: Kubernetes >Affects Versions: 3.0.0 >Reporter: Junjie Chen >Assignee: Junjie Chen >Priority: Minor > Fix For: 3.0.0 > > > 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 (v8.3.2#803003) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-28042) Support mapping spark.local.dir to hostPath volume
[ https://issues.apache.org/jira/browse/SPARK-28042?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16925265#comment-16925265 ] Jiaxin Shan commented on SPARK-28042: - [~vanzin] Do we usually backport these improvements to 2.4.x? > Support mapping spark.local.dir to hostPath volume > -- > > Key: SPARK-28042 > URL: https://issues.apache.org/jira/browse/SPARK-28042 > Project: Spark > Issue Type: Improvement > Components: Kubernetes >Affects Versions: 3.0.0 >Reporter: Junjie Chen >Assignee: Junjie Chen >Priority: Minor > Fix For: 3.0.0 > > > 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 (v8.3.2#803003) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
