[jira] [Commented] (YARN-8425) Yarn container getting killed due to running beyond physical memory limits
[ https://issues.apache.org/jira/browse/YARN-8425?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16512014#comment-16512014 ] Gour Saha commented on YARN-8425: - If you are not reporting a bug or improvement, sending email to u...@hadoop.apache.org is the right way to get your doubts and questions answered. Knowing your application's needs and asking for containers of the right size is the way to go. Disabling pmem check is not recommended in prod clusters. > Yarn container getting killed due to running beyond physical memory limits > -- > > Key: YARN-8425 > URL: https://issues.apache.org/jira/browse/YARN-8425 > Project: Hadoop YARN > Issue Type: Task > Components: applications, container-queuing, yarn >Affects Versions: 2.7.6 >Reporter: Tapas Sen >Priority: Major > Attachments: yarn_configuration_1.PNG, yarn_configuration_2.PNG, > yarn_configuration_3.PNG > > > Hi, > Getting these error. > > 2018-06-12 17:59:07,193 INFO [AsyncDispatcher event handler] > org.apache.hadoop.mapreduce.v2.app.job.impl.TaskAttemptImpl: Diagnostics > report from attempt_1527758146858_45040_m_08_3: Container > [pid=15498,containerID=container_e60_1527758146858_45040_01_41] is > running beyond physical memory limits. Current usage: 8.1 GB of 8 GB physical > memory used; 12.2 GB of 16.8 GB virtual memory used. Killing container. > > Yarn resource configuration will in attachment. > > Any lead would be appreciated. > > > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-8425) Yarn container getting killed due to running beyond physical memory limits
[ https://issues.apache.org/jira/browse/YARN-8425?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16512010#comment-16512010 ] Tapas Sen commented on YARN-8425: - So this is a prod cluster, what is the impact on the Cluster. Any standard documentation to do the same. > Yarn container getting killed due to running beyond physical memory limits > -- > > Key: YARN-8425 > URL: https://issues.apache.org/jira/browse/YARN-8425 > Project: Hadoop YARN > Issue Type: Task > Components: applications, container-queuing, yarn >Affects Versions: 2.7.6 >Reporter: Tapas Sen >Priority: Major > Attachments: yarn_configuration_1.PNG, yarn_configuration_2.PNG, > yarn_configuration_3.PNG > > > Hi, > Getting these error. > > 2018-06-12 17:59:07,193 INFO [AsyncDispatcher event handler] > org.apache.hadoop.mapreduce.v2.app.job.impl.TaskAttemptImpl: Diagnostics > report from attempt_1527758146858_45040_m_08_3: Container > [pid=15498,containerID=container_e60_1527758146858_45040_01_41] is > running beyond physical memory limits. Current usage: 8.1 GB of 8 GB physical > memory used; 12.2 GB of 16.8 GB virtual memory used. Killing container. > > Yarn resource configuration will in attachment. > > Any lead would be appreciated. > > > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-8425) Yarn container getting killed due to running beyond physical memory limits
[ https://issues.apache.org/jira/browse/YARN-8425?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16511241#comment-16511241 ] Gour Saha commented on YARN-8425: - If _yarn.nodemanager.pmem-check-enabled_ in your cluster is not explicitly set to false (since the default value is true) it is behaving as designed. Based on your requirement you can either request for containers higher than 8GB or set _yarn.nodemanager.pmem-check-enabled_ to false. > Yarn container getting killed due to running beyond physical memory limits > -- > > Key: YARN-8425 > URL: https://issues.apache.org/jira/browse/YARN-8425 > Project: Hadoop YARN > Issue Type: Task > Components: applications, container-queuing, yarn >Affects Versions: 2.7.6 >Reporter: Tapas Sen >Priority: Major > Attachments: yarn_configuration_1.PNG, yarn_configuration_2.PNG, > yarn_configuration_3.PNG > > > Hi, > Getting these error. > > 2018-06-12 17:59:07,193 INFO [AsyncDispatcher event handler] > org.apache.hadoop.mapreduce.v2.app.job.impl.TaskAttemptImpl: Diagnostics > report from attempt_1527758146858_45040_m_08_3: Container > [pid=15498,containerID=container_e60_1527758146858_45040_01_41] is > running beyond physical memory limits. Current usage: 8.1 GB of 8 GB physical > memory used; 12.2 GB of 16.8 GB virtual memory used. Killing container. > > Yarn resource configuration will in attachment. > > Any lead would be appreciated. > > > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
