[jira] [Resolved] (YARN-5517) Add GPU as a resource type for scheduling
[ https://issues.apache.org/jira/browse/YARN-5517?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jaeboo Jeong resolved YARN-5517. Resolution: Duplicate > Add GPU as a resource type for scheduling > - > > Key: YARN-5517 > URL: https://issues.apache.org/jira/browse/YARN-5517 > Project: Hadoop YARN > Issue Type: Improvement > Components: scheduler >Reporter: Jaeboo Jeong >Priority: Major > Attachments: RM-scheduler_metrics.jpg, YARN-5517-branch-2.7.1.patch, > aggregate_resource_allocation.jpg, container_example.jpg > > > Currently YARN only support scheduling based on memory and cpu. > There is the issue(YARN-3926) which proposed to extend the YARN resource > model. > And there is the issue(YARN-4122) to add support for GPU as a resource using > docker. > But these issues didn’t release yet so I just added GPU resource type like > memory and cpu. > I don’t consider GPU isolation like YARN-4122. > The properties for GPU resource type is similar to cpu core. > mapred-default.xml > mapreduce.map.gpu.cores (default 0) > mapreduce.reduce.gpu.cores(default 0) > yarn.app.mapreduce.am.resource.gpu-cores (default 0) > yarn-default.xml > yarn.scheduler.minimum-allocation-gcores (default 0) > yarn.scheduler.maximum-allocation-gcores (default 8) > yarn.nodemanager.resource.gcores (default 0) > I attached the patch for branch-2.7.1 -- 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] [Created] (YARN-5517) Add GPU as a resource type for scheduling on branch-2.7.1
Jaeboo Jeong created YARN-5517: -- Summary: Add GPU as a resource type for scheduling on branch-2.7.1 Key: YARN-5517 URL: https://issues.apache.org/jira/browse/YARN-5517 Project: Hadoop YARN Issue Type: Improvement Reporter: Jaeboo Jeong Currently YARN only support scheduling based on memory and cpu. There is the issue(YARN-3926) which proposed to extend the YARN resource model. And there is the issue(YARN-4122) to add support for GPU as a resource using docker. But these issues didn’t release yet so I just added GPU resource type like memory and cpu. I don’t consider GPU isolation like YARN-4122. The properties for GPU resource type is similar to cpu core. mapred-default.xml mapreduce.map.gpu.cores (default 0) mapreduce.reduce.gpu.cores (default 0) yarn.app.mapreduce.am.resource.gpu-cores (default 0) yarn-default.xml yarn.scheduler.minimum-allocation-gcores (default 0) yarn.scheduler.maximum-allocation-gcores (default 8) yarn.nodemanager.resource.gcores (default 0) I attached the patch for branch-2.7.1 -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Updated] (YARN-5517) Add GPU as a resource type for scheduling on branch-2.7.1
[ https://issues.apache.org/jira/browse/YARN-5517?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jaeboo Jeong updated YARN-5517: --- Attachment: RM-scheduler_metrics.jpg aggregate_resource_allocation.jpg container_example.jpg > Add GPU as a resource type for scheduling on branch-2.7.1 > - > > Key: YARN-5517 > URL: https://issues.apache.org/jira/browse/YARN-5517 > Project: Hadoop YARN > Issue Type: Improvement >Reporter: Jaeboo Jeong > Attachments: RM-scheduler_metrics.jpg, YARN-5517-branch-2.7.1.patch, > aggregate_resource_allocation.jpg, container_example.jpg > > > Currently YARN only support scheduling based on memory and cpu. > There is the issue(YARN-3926) which proposed to extend the YARN resource > model. > And there is the issue(YARN-4122) to add support for GPU as a resource using > docker. > But these issues didn’t release yet so I just added GPU resource type like > memory and cpu. > I don’t consider GPU isolation like YARN-4122. > The properties for GPU resource type is similar to cpu core. > mapred-default.xml > mapreduce.map.gpu.cores (default 0) > mapreduce.reduce.gpu.cores(default 0) > yarn.app.mapreduce.am.resource.gpu-cores (default 0) > yarn-default.xml > yarn.scheduler.minimum-allocation-gcores (default 0) > yarn.scheduler.maximum-allocation-gcores (default 8) > yarn.nodemanager.resource.gcores (default 0) > I attached the patch for branch-2.7.1 -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Updated] (YARN-5517) Add GPU as a resource type for scheduling on branch-2.7.1
[ https://issues.apache.org/jira/browse/YARN-5517?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jaeboo Jeong updated YARN-5517: --- Attachment: YARN-5517-branch-2.7.1.patch > Add GPU as a resource type for scheduling on branch-2.7.1 > - > > Key: YARN-5517 > URL: https://issues.apache.org/jira/browse/YARN-5517 > Project: Hadoop YARN > Issue Type: Improvement >Reporter: Jaeboo Jeong > Attachments: YARN-5517-branch-2.7.1.patch > > > Currently YARN only support scheduling based on memory and cpu. > There is the issue(YARN-3926) which proposed to extend the YARN resource > model. > And there is the issue(YARN-4122) to add support for GPU as a resource using > docker. > But these issues didn’t release yet so I just added GPU resource type like > memory and cpu. > I don’t consider GPU isolation like YARN-4122. > The properties for GPU resource type is similar to cpu core. > mapred-default.xml > mapreduce.map.gpu.cores (default 0) > mapreduce.reduce.gpu.cores(default 0) > yarn.app.mapreduce.am.resource.gpu-cores (default 0) > yarn-default.xml > yarn.scheduler.minimum-allocation-gcores (default 0) > yarn.scheduler.maximum-allocation-gcores (default 8) > yarn.nodemanager.resource.gcores (default 0) > I attached the patch for branch-2.7.1 -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Created] (YARN-5578) New network outbound configuration is not applied when NM starts
Jaeboo Jeong created YARN-5578: -- Summary: New network outbound configuration is not applied when NM starts Key: YARN-5578 URL: https://issues.apache.org/jira/browse/YARN-5578 Project: Hadoop YARN Issue Type: Bug Reporter: Jaeboo Jeong Priority: Minor Attachments: YARN-5578.patch I'm testing a network outbound bandwidth shaping by applying YARN-3366. If yarn.nodemanager.recovery.enabled is true, new configuration is not applied when NM starts. So network outbound bandwidth is always initial setting value. I just added the code to update states using existing container classes list when NM starts. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Updated] (YARN-5578) New network outbound configuration is not applied when NM starts
[ https://issues.apache.org/jira/browse/YARN-5578?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jaeboo Jeong updated YARN-5578: --- Attachment: YARN-5578.patch > New network outbound configuration is not applied when NM starts > > > Key: YARN-5578 > URL: https://issues.apache.org/jira/browse/YARN-5578 > Project: Hadoop YARN > Issue Type: Bug >Reporter: Jaeboo Jeong >Priority: Minor > Attachments: YARN-5578.patch > > > I'm testing a network outbound bandwidth shaping by applying YARN-3366. > If yarn.nodemanager.recovery.enabled is true, new configuration is not > applied when NM starts. > So network outbound bandwidth is always initial setting value. > I just added the code to update states using existing container classes list > when NM starts. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Updated] (YARN-6495) check docker container's exit code when writing to cgroup task files
[ https://issues.apache.org/jira/browse/YARN-6495?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jaeboo Jeong updated YARN-6495: --- Attachment: YARN-6495.003.patch > check docker container's exit code when writing to cgroup task files > > > Key: YARN-6495 > URL: https://issues.apache.org/jira/browse/YARN-6495 > Project: Hadoop YARN > Issue Type: Sub-task > Components: nodemanager >Reporter: Jaeboo Jeong >Assignee: Jim Brennan >Priority: Major > Labels: Docker > Attachments: YARN-6495.001.patch, YARN-6495.002.patch, > YARN-6495.003.patch > > > If I execute simple command like date on docker container, the application > failed to complete successfully. > for example, > {code} > $ yarn jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -shell_env YARN_CONTAINER_RUNTIME_TYPE=docker -shell_env > YARN_CONTAINER_RUNTIME_DOCKER_IMAGE=hadoop-docker -shell_command "date" -jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -num_containers 1 -timeout 360 > … > 17/04/12 00:16:40 INFO distributedshell.Client: Application did finished > unsuccessfully. YarnState=FINISHED, DSFinalStatus=FAILED. Breaking monitoring > loop > 17/04/12 00:16:40 ERROR distributedshell.Client: Application failed to > complete successfully > {code} > The error log is like below. > {code} > ... > Failed to write pid to file > /cgroup_parent/cpu/hadoop-yarn/container_/tasks - No such process > ... > {code} > When writing pid to cgroup tasks, container-executor doesn’t check docker > container’s status. > If the container finished very quickly, we can’t write pid to cgroup tasks, > and it is not problem. > So container-executor needs to check docker container’s exit code during > writing pid to cgroup tasks. -- 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-6495) check docker container's exit code when writing to cgroup task files
[ https://issues.apache.org/jira/browse/YARN-6495?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16584740#comment-16584740 ] Jaeboo Jeong commented on YARN-6495: [~Jim_Brennan] Thank you for sharing issue. I checked that and I agree with your opinion. Please close this issue. Anyway I just uploaded the patch for fixing previous patch. > check docker container's exit code when writing to cgroup task files > > > Key: YARN-6495 > URL: https://issues.apache.org/jira/browse/YARN-6495 > Project: Hadoop YARN > Issue Type: Sub-task > Components: nodemanager >Reporter: Jaeboo Jeong >Assignee: Jim Brennan >Priority: Major > Labels: Docker > Attachments: YARN-6495.001.patch, YARN-6495.002.patch, > YARN-6495.003.patch > > > If I execute simple command like date on docker container, the application > failed to complete successfully. > for example, > {code} > $ yarn jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -shell_env YARN_CONTAINER_RUNTIME_TYPE=docker -shell_env > YARN_CONTAINER_RUNTIME_DOCKER_IMAGE=hadoop-docker -shell_command "date" -jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -num_containers 1 -timeout 360 > … > 17/04/12 00:16:40 INFO distributedshell.Client: Application did finished > unsuccessfully. YarnState=FINISHED, DSFinalStatus=FAILED. Breaking monitoring > loop > 17/04/12 00:16:40 ERROR distributedshell.Client: Application failed to > complete successfully > {code} > The error log is like below. > {code} > ... > Failed to write pid to file > /cgroup_parent/cpu/hadoop-yarn/container_/tasks - No such process > ... > {code} > When writing pid to cgroup tasks, container-executor doesn’t check docker > container’s status. > If the container finished very quickly, we can’t write pid to cgroup tasks, > and it is not problem. > So container-executor needs to check docker container’s exit code during > writing pid to cgroup tasks. -- 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] [Updated] (YARN-6495) check docker container's exit code when writing to cgroup task files
[ https://issues.apache.org/jira/browse/YARN-6495?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jaeboo Jeong updated YARN-6495: --- Attachment: (was: YARN-6495.003.patch) > check docker container's exit code when writing to cgroup task files > > > Key: YARN-6495 > URL: https://issues.apache.org/jira/browse/YARN-6495 > Project: Hadoop YARN > Issue Type: Sub-task > Components: nodemanager >Reporter: Jaeboo Jeong >Assignee: Jim Brennan >Priority: Major > Labels: Docker > Attachments: YARN-6495.001.patch, YARN-6495.002.patch > > > If I execute simple command like date on docker container, the application > failed to complete successfully. > for example, > {code} > $ yarn jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -shell_env YARN_CONTAINER_RUNTIME_TYPE=docker -shell_env > YARN_CONTAINER_RUNTIME_DOCKER_IMAGE=hadoop-docker -shell_command "date" -jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -num_containers 1 -timeout 360 > … > 17/04/12 00:16:40 INFO distributedshell.Client: Application did finished > unsuccessfully. YarnState=FINISHED, DSFinalStatus=FAILED. Breaking monitoring > loop > 17/04/12 00:16:40 ERROR distributedshell.Client: Application failed to > complete successfully > {code} > The error log is like below. > {code} > ... > Failed to write pid to file > /cgroup_parent/cpu/hadoop-yarn/container_/tasks - No such process > ... > {code} > When writing pid to cgroup tasks, container-executor doesn’t check docker > container’s status. > If the container finished very quickly, we can’t write pid to cgroup tasks, > and it is not problem. > So container-executor needs to check docker container’s exit code during > writing pid to cgroup tasks. -- 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] [Updated] (YARN-6495) check docker container's exit code when writing to cgroup task files
[ https://issues.apache.org/jira/browse/YARN-6495?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jaeboo Jeong updated YARN-6495: --- Attachment: YARN-6495.003.patch > check docker container's exit code when writing to cgroup task files > > > Key: YARN-6495 > URL: https://issues.apache.org/jira/browse/YARN-6495 > Project: Hadoop YARN > Issue Type: Sub-task > Components: nodemanager >Reporter: Jaeboo Jeong >Assignee: Jim Brennan >Priority: Major > Labels: Docker > Attachments: YARN-6495.001.patch, YARN-6495.002.patch, > YARN-6495.003.patch > > > If I execute simple command like date on docker container, the application > failed to complete successfully. > for example, > {code} > $ yarn jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -shell_env YARN_CONTAINER_RUNTIME_TYPE=docker -shell_env > YARN_CONTAINER_RUNTIME_DOCKER_IMAGE=hadoop-docker -shell_command "date" -jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -num_containers 1 -timeout 360 > … > 17/04/12 00:16:40 INFO distributedshell.Client: Application did finished > unsuccessfully. YarnState=FINISHED, DSFinalStatus=FAILED. Breaking monitoring > loop > 17/04/12 00:16:40 ERROR distributedshell.Client: Application failed to > complete successfully > {code} > The error log is like below. > {code} > ... > Failed to write pid to file > /cgroup_parent/cpu/hadoop-yarn/container_/tasks - No such process > ... > {code} > When writing pid to cgroup tasks, container-executor doesn’t check docker > container’s status. > If the container finished very quickly, we can’t write pid to cgroup tasks, > and it is not problem. > So container-executor needs to check docker container’s exit code during > writing pid to cgroup tasks. -- 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-6494) add mounting of HDFS Short-Circuit path for docker containers
[ https://issues.apache.org/jira/browse/YARN-6494?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16107424#comment-16107424 ] Jaeboo Jeong commented on YARN-6494: Thank you for the feedbacks. [~dan...@cloudera.com] Current patch was referenced another test cases in TestDockerContainerRuntime. DockerLinuxContainerRuntime write the docker command to a temporary file, so I think it's not bad to compare command strings with the contents of the temporary file. Of course, it is true that it is difficult to memorize the order of commands and make sure that they are correct. [~shaneku...@gmail.com] I agree with the white list for safe volume mounts. However I thought this case is different from ordinary file mounts. I think that the path for the short circuit is the configuration of the cluster, and the cluster user doesn't need to consider about that path. I just think it's just part of the cluster setup. > add mounting of HDFS Short-Circuit path for docker containers > - > > Key: YARN-6494 > URL: https://issues.apache.org/jira/browse/YARN-6494 > Project: Hadoop YARN > Issue Type: Sub-task > Components: nodemanager >Reporter: Jaeboo Jeong >Assignee: Jaeboo Jeong > Attachments: YARN-6494.001.patch, YARN-6494.002.patch > > > Currently there is a error message about HDFS short-circuit when docker > container start. > {code} > WARN [main] org.apache.hadoop.hdfs.shortcircuit.DomainSocketFactory: error > creating DomainSocket > java.net.ConnectException: connect(2) error: No such file or directory when > trying to connect to ‘xxx’ > at org.apache.hadoop.net.unix.DomainSocket.connect0(Native Method) > at org.apache.hadoop.net.unix.DomainSocket.connect(DomainSocket.java:250) > at > org.apache.hadoop.hdfs.shortcircuit.DomainSocketFactory.createSocket(DomainSocketFactory.java:164) > at > org.apache.hadoop.hdfs.BlockReaderFactory.nextDomainPeer(BlockReaderFactory.java:752) > ... > {code} > if dfs.client.read.shortcircuit is true and dfs.domain.socket.path isn't > equal “”, we need to mount volume for short-circuit path. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-6494) add mounting of HDFS Short-Circuit path for docker containers
[ https://issues.apache.org/jira/browse/YARN-6494?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16112098#comment-16112098 ] Jaeboo Jeong commented on YARN-6494: If we use white list for all kinds of mounting, the cluster administrator has to be concerned with additional settings for the docker environment. And there is no way to know what configuration is missed. I'm not sure it is fine, because after we finished the cluster setup, we expect all the containers in the cluster to run in the same environment. However, I think it is safe for the cluster administrator to be more careful because administrator does not know which applications are running. I think it would be better to document this later. So I agree to close this issue. > add mounting of HDFS Short-Circuit path for docker containers > - > > Key: YARN-6494 > URL: https://issues.apache.org/jira/browse/YARN-6494 > Project: Hadoop YARN > Issue Type: Sub-task > Components: nodemanager >Reporter: Jaeboo Jeong >Assignee: Jaeboo Jeong > Attachments: YARN-6494.001.patch, YARN-6494.002.patch > > > Currently there is a error message about HDFS short-circuit when docker > container start. > {code} > WARN [main] org.apache.hadoop.hdfs.shortcircuit.DomainSocketFactory: error > creating DomainSocket > java.net.ConnectException: connect(2) error: No such file or directory when > trying to connect to ‘xxx’ > at org.apache.hadoop.net.unix.DomainSocket.connect0(Native Method) > at org.apache.hadoop.net.unix.DomainSocket.connect(DomainSocket.java:250) > at > org.apache.hadoop.hdfs.shortcircuit.DomainSocketFactory.createSocket(DomainSocketFactory.java:164) > at > org.apache.hadoop.hdfs.BlockReaderFactory.nextDomainPeer(BlockReaderFactory.java:752) > ... > {code} > if dfs.client.read.shortcircuit is true and dfs.domain.socket.path isn't > equal “”, we need to mount volume for short-circuit path. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-6495) check docker container's exit code when writing to cgroup task files
[ https://issues.apache.org/jira/browse/YARN-6495?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16412333#comment-16412333 ] Jaeboo Jeong commented on YARN-6495: It seems to be a problem because the execution of the docker command and the writing of the cgroup are executed independently. However, since both tasks are not independent, I think it would be better to check the command exit code during writing cgroup. > check docker container's exit code when writing to cgroup task files > > > Key: YARN-6495 > URL: https://issues.apache.org/jira/browse/YARN-6495 > Project: Hadoop YARN > Issue Type: Sub-task > Components: nodemanager >Reporter: Jaeboo Jeong >Assignee: Jaeboo Jeong >Priority: Major > Attachments: YARN-6495.001.patch > > > If I execute simple command like date on docker container, the application > failed to complete successfully. > for example, > {code} > $ yarn jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -shell_env YARN_CONTAINER_RUNTIME_TYPE=docker -shell_env > YARN_CONTAINER_RUNTIME_DOCKER_IMAGE=hadoop-docker -shell_command "date" -jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -num_containers 1 -timeout 360 > … > 17/04/12 00:16:40 INFO distributedshell.Client: Application did finished > unsuccessfully. YarnState=FINISHED, DSFinalStatus=FAILED. Breaking monitoring > loop > 17/04/12 00:16:40 ERROR distributedshell.Client: Application failed to > complete successfully > {code} > The error log is like below. > {code} > ... > Failed to write pid to file > /cgroup_parent/cpu/hadoop-yarn/container_/tasks - No such process > ... > {code} > When writing pid to cgroup tasks, container-executor doesn’t check docker > container’s status. > If the container finished very quickly, we can’t write pid to cgroup tasks, > and it is not problem. > So container-executor needs to check docker container’s exit code during > writing pid to cgroup tasks. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Updated] (YARN-6434) When setting environment variables, can't use comma for a list of value in key = value pairs.
[ https://issues.apache.org/jira/browse/YARN-6434?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jaeboo Jeong updated YARN-6434: --- Attachment: YARN-6434-trunk.001.patch > When setting environment variables, can't use comma for a list of value in > key = value pairs. > - > > Key: YARN-6434 > URL: https://issues.apache.org/jira/browse/YARN-6434 > Project: Hadoop YARN > Issue Type: Improvement >Reporter: Jaeboo Jeong >Priority: Major > Attachments: YARN-6434-trunk.001.patch, YARN-6434.001.patch > > > We can set environment variables using yarn.app.mapreduce.am.env, > mapreduce.map.env, mapreduce.reduce.env. > There is no problem if we use key=value pairs like X=Y, X=$Y. > However If we want to set key=a list of value pair(e.g. X=Y,Z), we can’t. > This is related to YARN-4595. > The attached patch is based on YARN-3768. > We can set environment variables like below. > {code} > mapreduce.map.env="YARN_CONTAINER_RUNTIME_TYPE=docker,YARN_CONTAINER_RUNTIME_DOCKER_IMAGE=hadoop-docker,YARN_CONTAINER_RUNTIME_DOCKER_LOCAL_RESOURCE_MOUNTS=\"/dir1:/targetdir1,/dir2:/targetdir2\"" > {code} -- 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-6495) check docker container's exit code when writing to cgroup task files
[ https://issues.apache.org/jira/browse/YARN-6495?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16419118#comment-16419118 ] Jaeboo Jeong commented on YARN-6495: In write_pid_to_cgroup_as_root(), different error cases are already being checked except docker command status. > check docker container's exit code when writing to cgroup task files > > > Key: YARN-6495 > URL: https://issues.apache.org/jira/browse/YARN-6495 > Project: Hadoop YARN > Issue Type: Sub-task > Components: nodemanager >Reporter: Jaeboo Jeong >Assignee: Jaeboo Jeong >Priority: Major > Attachments: YARN-6495.001.patch > > > If I execute simple command like date on docker container, the application > failed to complete successfully. > for example, > {code} > $ yarn jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -shell_env YARN_CONTAINER_RUNTIME_TYPE=docker -shell_env > YARN_CONTAINER_RUNTIME_DOCKER_IMAGE=hadoop-docker -shell_command "date" -jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -num_containers 1 -timeout 360 > … > 17/04/12 00:16:40 INFO distributedshell.Client: Application did finished > unsuccessfully. YarnState=FINISHED, DSFinalStatus=FAILED. Breaking monitoring > loop > 17/04/12 00:16:40 ERROR distributedshell.Client: Application failed to > complete successfully > {code} > The error log is like below. > {code} > ... > Failed to write pid to file > /cgroup_parent/cpu/hadoop-yarn/container_/tasks - No such process > ... > {code} > When writing pid to cgroup tasks, container-executor doesn’t check docker > container’s status. > If the container finished very quickly, we can’t write pid to cgroup tasks, > and it is not problem. > So container-executor needs to check docker container’s exit code during > writing pid to cgroup tasks. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-6434) When setting environment variables, can't use comma for a list of value in key = value pairs.
[ https://issues.apache.org/jira/browse/YARN-6434?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16436615#comment-16436615 ] Jaeboo Jeong commented on YARN-6434: Jim Brennan, Thank you for informing me. I can see variables clearly, so I think it is good for readability. > When setting environment variables, can't use comma for a list of value in > key = value pairs. > - > > Key: YARN-6434 > URL: https://issues.apache.org/jira/browse/YARN-6434 > Project: Hadoop YARN > Issue Type: Improvement >Reporter: Jaeboo Jeong >Priority: Major > Attachments: YARN-6434-trunk.001.patch, YARN-6434.001.patch > > > We can set environment variables using yarn.app.mapreduce.am.env, > mapreduce.map.env, mapreduce.reduce.env. > There is no problem if we use key=value pairs like X=Y, X=$Y. > However If we want to set key=a list of value pair(e.g. X=Y,Z), we can’t. > This is related to YARN-4595. > The attached patch is based on YARN-3768. > We can set environment variables like below. > {code} > mapreduce.map.env="YARN_CONTAINER_RUNTIME_TYPE=docker,YARN_CONTAINER_RUNTIME_DOCKER_IMAGE=hadoop-docker,YARN_CONTAINER_RUNTIME_DOCKER_LOCAL_RESOURCE_MOUNTS=\"/dir1:/targetdir1,/dir2:/targetdir2\"" > {code} -- 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] [Updated] (YARN-6495) check docker container's exit code when writing to cgroup task files
[ https://issues.apache.org/jira/browse/YARN-6495?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jaeboo Jeong updated YARN-6495: --- Attachment: YARN-6495.002.patch > check docker container's exit code when writing to cgroup task files > > > Key: YARN-6495 > URL: https://issues.apache.org/jira/browse/YARN-6495 > Project: Hadoop YARN > Issue Type: Sub-task > Components: nodemanager >Reporter: Jaeboo Jeong >Assignee: Jaeboo Jeong >Priority: Major > Attachments: YARN-6495.001.patch, YARN-6495.002.patch > > > If I execute simple command like date on docker container, the application > failed to complete successfully. > for example, > {code} > $ yarn jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -shell_env YARN_CONTAINER_RUNTIME_TYPE=docker -shell_env > YARN_CONTAINER_RUNTIME_DOCKER_IMAGE=hadoop-docker -shell_command "date" -jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -num_containers 1 -timeout 360 > … > 17/04/12 00:16:40 INFO distributedshell.Client: Application did finished > unsuccessfully. YarnState=FINISHED, DSFinalStatus=FAILED. Breaking monitoring > loop > 17/04/12 00:16:40 ERROR distributedshell.Client: Application failed to > complete successfully > {code} > The error log is like below. > {code} > ... > Failed to write pid to file > /cgroup_parent/cpu/hadoop-yarn/container_/tasks - No such process > ... > {code} > When writing pid to cgroup tasks, container-executor doesn’t check docker > container’s status. > If the container finished very quickly, we can’t write pid to cgroup tasks, > and it is not problem. > So container-executor needs to check docker container’s exit code during > writing pid to cgroup tasks. -- 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-6495) check docker container's exit code when writing to cgroup task files
[ https://issues.apache.org/jira/browse/YARN-6495?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16437403#comment-16437403 ] Jaeboo Jeong commented on YARN-6495: I separated docker container's exit code from cgroup failed exit code. And I uploaded the patch file. > check docker container's exit code when writing to cgroup task files > > > Key: YARN-6495 > URL: https://issues.apache.org/jira/browse/YARN-6495 > Project: Hadoop YARN > Issue Type: Sub-task > Components: nodemanager >Reporter: Jaeboo Jeong >Assignee: Jaeboo Jeong >Priority: Major > Attachments: YARN-6495.001.patch, YARN-6495.002.patch > > > If I execute simple command like date on docker container, the application > failed to complete successfully. > for example, > {code} > $ yarn jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -shell_env YARN_CONTAINER_RUNTIME_TYPE=docker -shell_env > YARN_CONTAINER_RUNTIME_DOCKER_IMAGE=hadoop-docker -shell_command "date" -jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -num_containers 1 -timeout 360 > … > 17/04/12 00:16:40 INFO distributedshell.Client: Application did finished > unsuccessfully. YarnState=FINISHED, DSFinalStatus=FAILED. Breaking monitoring > loop > 17/04/12 00:16:40 ERROR distributedshell.Client: Application failed to > complete successfully > {code} > The error log is like below. > {code} > ... > Failed to write pid to file > /cgroup_parent/cpu/hadoop-yarn/container_/tasks - No such process > ... > {code} > When writing pid to cgroup tasks, container-executor doesn’t check docker > container’s status. > If the container finished very quickly, we can’t write pid to cgroup tasks, > and it is not problem. > So container-executor needs to check docker container’s exit code during > writing pid to cgroup tasks. -- 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-6495) check docker container's exit code when writing to cgroup task files
[ https://issues.apache.org/jira/browse/YARN-6495?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16447062#comment-16447062 ] Jaeboo Jeong commented on YARN-6495: [~ebadger], Oops, the patch(002 patch) is wrong. I will fix it on trunk. And what if you think about the two topics(this issue and improvement of write_pid_to_cgroup_as_root()) separately? As you mentioned, write_pid_to_cgroup_as_root() doesn't differentiat exit codes but it is not the reason of this issue. > check docker container's exit code when writing to cgroup task files > > > Key: YARN-6495 > URL: https://issues.apache.org/jira/browse/YARN-6495 > Project: Hadoop YARN > Issue Type: Sub-task > Components: nodemanager >Reporter: Jaeboo Jeong >Assignee: Jaeboo Jeong >Priority: Major > Attachments: YARN-6495.001.patch, YARN-6495.002.patch > > > If I execute simple command like date on docker container, the application > failed to complete successfully. > for example, > {code} > $ yarn jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -shell_env YARN_CONTAINER_RUNTIME_TYPE=docker -shell_env > YARN_CONTAINER_RUNTIME_DOCKER_IMAGE=hadoop-docker -shell_command "date" -jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -num_containers 1 -timeout 360 > … > 17/04/12 00:16:40 INFO distributedshell.Client: Application did finished > unsuccessfully. YarnState=FINISHED, DSFinalStatus=FAILED. Breaking monitoring > loop > 17/04/12 00:16:40 ERROR distributedshell.Client: Application failed to > complete successfully > {code} > The error log is like below. > {code} > ... > Failed to write pid to file > /cgroup_parent/cpu/hadoop-yarn/container_/tasks - No such process > ... > {code} > When writing pid to cgroup tasks, container-executor doesn’t check docker > container’s status. > If the container finished very quickly, we can’t write pid to cgroup tasks, > and it is not problem. > So container-executor needs to check docker container’s exit code during > writing pid to cgroup tasks. -- 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-5517) Add GPU as a resource type for scheduling
[ https://issues.apache.org/jira/browse/YARN-5517?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15884747#comment-15884747 ] Jaeboo Jeong commented on YARN-5517: This patch just is intended for use with hadoop 2.7. I agree with Vinod's comment. > Add GPU as a resource type for scheduling > - > > Key: YARN-5517 > URL: https://issues.apache.org/jira/browse/YARN-5517 > Project: Hadoop YARN > Issue Type: Improvement > Components: scheduler >Reporter: Jaeboo Jeong > Attachments: aggregate_resource_allocation.jpg, > container_example.jpg, RM-scheduler_metrics.jpg, YARN-5517-branch-2.7.1.patch > > > Currently YARN only support scheduling based on memory and cpu. > There is the issue(YARN-3926) which proposed to extend the YARN resource > model. > And there is the issue(YARN-4122) to add support for GPU as a resource using > docker. > But these issues didn’t release yet so I just added GPU resource type like > memory and cpu. > I don’t consider GPU isolation like YARN-4122. > The properties for GPU resource type is similar to cpu core. > mapred-default.xml > mapreduce.map.gpu.cores (default 0) > mapreduce.reduce.gpu.cores(default 0) > yarn.app.mapreduce.am.resource.gpu-cores (default 0) > yarn-default.xml > yarn.scheduler.minimum-allocation-gcores (default 0) > yarn.scheduler.maximum-allocation-gcores (default 8) > yarn.nodemanager.resource.gcores (default 0) > I attached the patch for branch-2.7.1 -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-4595) Add support for configurable read-only mounts when launching Docker containers
[ https://issues.apache.org/jira/browse/YARN-4595?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15942914#comment-15942914 ] Jaeboo Jeong commented on YARN-4595: If I want to use multiple mounts(e.g. /dir1:/targerdir1,/dir2:/targetdir2), it couldn't assign variables correctly. Because during setting environment variables from input string(at Apps.java), the comma is used for separator. example {code}mapreduce.map.env="YARN_CONTAINER_RUNTIME_TYPE=docker,YARN_CONTAINER_RUNTIME_DOCKER_IMAGE=hadoop-docker,YARN_CONTAINER_RUNTIME_DOCKER_LOCAL_RESOURCE_MOUNTS=/dir1:/targetdir1,/dir2:/targetdir2"{code} > Add support for configurable read-only mounts when launching Docker containers > -- > > Key: YARN-4595 > URL: https://issues.apache.org/jira/browse/YARN-4595 > Project: Hadoop YARN > Issue Type: Sub-task > Components: yarn >Reporter: Billie Rinaldi >Assignee: Billie Rinaldi > Fix For: 2.9.0, 3.0.0-alpha1 > > Attachments: YARN-4595.1.patch, YARN-4595.2.patch, YARN-4595.3.patch, > YARN-4595.4.patch, YARN-4595.5.patch > > > Mounting files or directories from the host is one way of passing > configuration and other information into a docker container. We could allow > the user to set a list of mounts in the environment of ContainerLaunchContext > (e.g. /dir1:/targetdir1,/dir2:/targetdir2). These would be mounted read-only > to the specified target locations. > Due to permissions and user concerns, for this ticket we will require the > mounts to be resources that are in the distributed cache. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Created] (YARN-6434) When setting environment variables, can't use comma for a list of value in key = value pairs.
Jaeboo Jeong created YARN-6434: -- Summary: When setting environment variables, can't use comma for a list of value in key = value pairs. Key: YARN-6434 URL: https://issues.apache.org/jira/browse/YARN-6434 Project: Hadoop YARN Issue Type: Improvement Reporter: Jaeboo Jeong We can set environment variables using yarn.app.mapreduce.am.env, mapreduce.map.env, mapreduce.reduce.env. There is no problem if we use key=value pairs like X=Y, X=$Y. However If we want to set key=a list of value pair(e.g. X=Y,Z), we can’t. This is related to YARN-4595. The attached patch is based on YARN-3768. We can set environment variables like below. {code} mapreduce.map.env="YARN_CONTAINER_RUNTIME_TYPE=docker,YARN_CONTAINER_RUNTIME_DOCKER_IMAGE=hadoop-docker,YARN_CONTAINER_RUNTIME_DOCKER_LOCAL_RESOURCE_MOUNTS=\"/dir1:/targetdir1,/dir2:/targetdir2\"" {code} -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Updated] (YARN-6434) When setting environment variables, can't use comma for a list of value in key = value pairs.
[ https://issues.apache.org/jira/browse/YARN-6434?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jaeboo Jeong updated YARN-6434: --- Attachment: YARN-6434.001.patch > When setting environment variables, can't use comma for a list of value in > key = value pairs. > - > > Key: YARN-6434 > URL: https://issues.apache.org/jira/browse/YARN-6434 > Project: Hadoop YARN > Issue Type: Improvement >Reporter: Jaeboo Jeong > Attachments: YARN-6434.001.patch > > > We can set environment variables using yarn.app.mapreduce.am.env, > mapreduce.map.env, mapreduce.reduce.env. > There is no problem if we use key=value pairs like X=Y, X=$Y. > However If we want to set key=a list of value pair(e.g. X=Y,Z), we can’t. > This is related to YARN-4595. > The attached patch is based on YARN-3768. > We can set environment variables like below. > {code} > mapreduce.map.env="YARN_CONTAINER_RUNTIME_TYPE=docker,YARN_CONTAINER_RUNTIME_DOCKER_IMAGE=hadoop-docker,YARN_CONTAINER_RUNTIME_DOCKER_LOCAL_RESOURCE_MOUNTS=\"/dir1:/targetdir1,/dir2:/targetdir2\"" > {code} -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Created] (YARN-6454) Add support for setting environment variables for docker containers
Jaeboo Jeong created YARN-6454: -- Summary: Add support for setting environment variables for docker containers Key: YARN-6454 URL: https://issues.apache.org/jira/browse/YARN-6454 Project: Hadoop YARN Issue Type: Improvement Components: nodemanager Reporter: Jaeboo Jeong Docker allows to set environment variables for your containers with the -e flag. You can set environment variables like below. {code} YARN_CONTAINER_RUNTIME_DOCKER_ENVIRONMENT_VARIABLES=“HADOOP_CONF_DIR=$HADOOP_CONF_DIR,HADOOP_HDFS_HOME=/opt/hadoop" {code} If you want to set a list of values, apply YARN-6434 first. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Updated] (YARN-6454) Add support for setting environment variables for docker containers
[ https://issues.apache.org/jira/browse/YARN-6454?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jaeboo Jeong updated YARN-6454: --- Attachment: YARN-6454.001.patch > Add support for setting environment variables for docker containers > --- > > Key: YARN-6454 > URL: https://issues.apache.org/jira/browse/YARN-6454 > Project: Hadoop YARN > Issue Type: Improvement > Components: nodemanager >Reporter: Jaeboo Jeong > Attachments: YARN-6454.001.patch > > > Docker allows to set environment variables for your containers with the -e > flag. > You can set environment variables like below. > {code} > YARN_CONTAINER_RUNTIME_DOCKER_ENVIRONMENT_VARIABLES=“HADOOP_CONF_DIR=$HADOOP_CONF_DIR,HADOOP_HDFS_HOME=/opt/hadoop" > {code} > If you want to set a list of values, apply YARN-6434 first. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Created] (YARN-6494) add mounting of HDFS Short-Circuit path for docker containers
Jaeboo Jeong created YARN-6494: -- Summary: add mounting of HDFS Short-Circuit path for docker containers Key: YARN-6494 URL: https://issues.apache.org/jira/browse/YARN-6494 Project: Hadoop YARN Issue Type: Improvement Components: nodemanager Reporter: Jaeboo Jeong Currently there is a error message about HDFS short-circuit when docker container start. {code} WARN [main] org.apache.hadoop.hdfs.shortcircuit.DomainSocketFactory: error creating DomainSocket java.net.ConnectException: connect(2) error: No such file or directory when trying to connect to ‘xxx’ at org.apache.hadoop.net.unix.DomainSocket.connect0(Native Method) at org.apache.hadoop.net.unix.DomainSocket.connect(DomainSocket.java:250) at org.apache.hadoop.hdfs.shortcircuit.DomainSocketFactory.createSocket(DomainSocketFactory.java:164) at org.apache.hadoop.hdfs.BlockReaderFactory.nextDomainPeer(BlockReaderFactory.java:752) ... {code} if dfs.client.read.shortcircuit is true and dfs.domain.socket.path isn't equal “”, we need to mount volume for short-circuit path. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Updated] (YARN-6494) add mounting of HDFS Short-Circuit path for docker containers
[ https://issues.apache.org/jira/browse/YARN-6494?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jaeboo Jeong updated YARN-6494: --- Attachment: YARN-6494.001.patch > add mounting of HDFS Short-Circuit path for docker containers > - > > Key: YARN-6494 > URL: https://issues.apache.org/jira/browse/YARN-6494 > Project: Hadoop YARN > Issue Type: Improvement > Components: nodemanager >Reporter: Jaeboo Jeong > Attachments: YARN-6494.001.patch > > > Currently there is a error message about HDFS short-circuit when docker > container start. > {code} > WARN [main] org.apache.hadoop.hdfs.shortcircuit.DomainSocketFactory: error > creating DomainSocket > java.net.ConnectException: connect(2) error: No such file or directory when > trying to connect to ‘xxx’ > at org.apache.hadoop.net.unix.DomainSocket.connect0(Native Method) > at org.apache.hadoop.net.unix.DomainSocket.connect(DomainSocket.java:250) > at > org.apache.hadoop.hdfs.shortcircuit.DomainSocketFactory.createSocket(DomainSocketFactory.java:164) > at > org.apache.hadoop.hdfs.BlockReaderFactory.nextDomainPeer(BlockReaderFactory.java:752) > ... > {code} > if dfs.client.read.shortcircuit is true and dfs.domain.socket.path isn't > equal “”, we need to mount volume for short-circuit path. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Created] (YARN-6495) check docker container's exit code when writing to cgroup task files
Jaeboo Jeong created YARN-6495: -- Summary: check docker container's exit code when writing to cgroup task files Key: YARN-6495 URL: https://issues.apache.org/jira/browse/YARN-6495 Project: Hadoop YARN Issue Type: Improvement Components: nodemanager Reporter: Jaeboo Jeong If I execute simple command like date on docker container, the application failed to complete successfully. for example, {code} $ yarn jar $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar -shell_env YARN_CONTAINER_RUNTIME_TYPE=docker -shell_env YARN_CONTAINER_RUNTIME_DOCKER_IMAGE=hadoop-docker -shell_command "date" -jar $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar -num_containers 1 -timeout 360 … 17/04/12 00:16:40 INFO distributedshell.Client: Application did finished unsuccessfully. YarnState=FINISHED, DSFinalStatus=FAILED. Breaking monitoring loop 17/04/12 00:16:40 ERROR distributedshell.Client: Application failed to complete successfully {code} The error log is like below. {code} ... Failed to write pid to file /cgroup_parent/cpu/hadoop-yarn/container_/tasks - No such process ... {code} When writing pid to cgroup tasks, container-executor doesn’t check docker container’s status. If the container finished very quickly, we can’t write pid to cgroup tasks, and it is not problem. So container-executor needs to check docker container’s exit code during writing pid to cgroup tasks. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Updated] (YARN-6495) check docker container's exit code when writing to cgroup task files
[ https://issues.apache.org/jira/browse/YARN-6495?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jaeboo Jeong updated YARN-6495: --- Attachment: YARN-6495.001.patch > check docker container's exit code when writing to cgroup task files > > > Key: YARN-6495 > URL: https://issues.apache.org/jira/browse/YARN-6495 > Project: Hadoop YARN > Issue Type: Improvement > Components: nodemanager >Reporter: Jaeboo Jeong > Attachments: YARN-6495.001.patch > > > If I execute simple command like date on docker container, the application > failed to complete successfully. > for example, > {code} > $ yarn jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -shell_env YARN_CONTAINER_RUNTIME_TYPE=docker -shell_env > YARN_CONTAINER_RUNTIME_DOCKER_IMAGE=hadoop-docker -shell_command "date" -jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -num_containers 1 -timeout 360 > … > 17/04/12 00:16:40 INFO distributedshell.Client: Application did finished > unsuccessfully. YarnState=FINISHED, DSFinalStatus=FAILED. Breaking monitoring > loop > 17/04/12 00:16:40 ERROR distributedshell.Client: Application failed to > complete successfully > {code} > The error log is like below. > {code} > ... > Failed to write pid to file > /cgroup_parent/cpu/hadoop-yarn/container_/tasks - No such process > ... > {code} > When writing pid to cgroup tasks, container-executor doesn’t check docker > container’s status. > If the container finished very quickly, we can’t write pid to cgroup tasks, > and it is not problem. > So container-executor needs to check docker container’s exit code during > writing pid to cgroup tasks. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Updated] (YARN-6494) add mounting of HDFS Short-Circuit path for docker containers
[ https://issues.apache.org/jira/browse/YARN-6494?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jaeboo Jeong updated YARN-6494: --- Attachment: YARN-6494.002.patch > add mounting of HDFS Short-Circuit path for docker containers > - > > Key: YARN-6494 > URL: https://issues.apache.org/jira/browse/YARN-6494 > Project: Hadoop YARN > Issue Type: Improvement > Components: nodemanager >Reporter: Jaeboo Jeong >Assignee: Jaeboo Jeong > Attachments: YARN-6494.001.patch, YARN-6494.002.patch > > > Currently there is a error message about HDFS short-circuit when docker > container start. > {code} > WARN [main] org.apache.hadoop.hdfs.shortcircuit.DomainSocketFactory: error > creating DomainSocket > java.net.ConnectException: connect(2) error: No such file or directory when > trying to connect to ‘xxx’ > at org.apache.hadoop.net.unix.DomainSocket.connect0(Native Method) > at org.apache.hadoop.net.unix.DomainSocket.connect(DomainSocket.java:250) > at > org.apache.hadoop.hdfs.shortcircuit.DomainSocketFactory.createSocket(DomainSocketFactory.java:164) > at > org.apache.hadoop.hdfs.BlockReaderFactory.nextDomainPeer(BlockReaderFactory.java:752) > ... > {code} > if dfs.client.read.shortcircuit is true and dfs.domain.socket.path isn't > equal “”, we need to mount volume for short-circuit path. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-6494) add mounting of HDFS Short-Circuit path for docker containers
[ https://issues.apache.org/jira/browse/YARN-6494?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15978913#comment-15978913 ] Jaeboo Jeong commented on YARN-6494: On branch-3.0.0-alpha2, it compiled well. But on trunk branch, there is a compile error because of HDFS-11596. I made a new patch. > add mounting of HDFS Short-Circuit path for docker containers > - > > Key: YARN-6494 > URL: https://issues.apache.org/jira/browse/YARN-6494 > Project: Hadoop YARN > Issue Type: Improvement > Components: nodemanager >Reporter: Jaeboo Jeong >Assignee: Jaeboo Jeong > Attachments: YARN-6494.001.patch, YARN-6494.002.patch > > > Currently there is a error message about HDFS short-circuit when docker > container start. > {code} > WARN [main] org.apache.hadoop.hdfs.shortcircuit.DomainSocketFactory: error > creating DomainSocket > java.net.ConnectException: connect(2) error: No such file or directory when > trying to connect to ‘xxx’ > at org.apache.hadoop.net.unix.DomainSocket.connect0(Native Method) > at org.apache.hadoop.net.unix.DomainSocket.connect(DomainSocket.java:250) > at > org.apache.hadoop.hdfs.shortcircuit.DomainSocketFactory.createSocket(DomainSocketFactory.java:164) > at > org.apache.hadoop.hdfs.BlockReaderFactory.nextDomainPeer(BlockReaderFactory.java:752) > ... > {code} > if dfs.client.read.shortcircuit is true and dfs.domain.socket.path isn't > equal “”, we need to mount volume for short-circuit path. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-6495) check docker container's exit code when writing to cgroup task files
[ https://issues.apache.org/jira/browse/YARN-6495?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15983033#comment-15983033 ] Jaeboo Jeong commented on YARN-6495: I tested and developed on branch-2.7 because my cluster is running with 2.7. To use docker container in LCE, I patched all issues in YARN-3611. However I tested based on the branch-2 commit(3b7bb7b94b1974e74556e787e6bec7549040b3a5), the application failed to complete successfully too. I didn't do anything to getting failure. > check docker container's exit code when writing to cgroup task files > > > Key: YARN-6495 > URL: https://issues.apache.org/jira/browse/YARN-6495 > Project: Hadoop YARN > Issue Type: Improvement > Components: nodemanager >Reporter: Jaeboo Jeong >Assignee: Jaeboo Jeong > Attachments: YARN-6495.001.patch > > > If I execute simple command like date on docker container, the application > failed to complete successfully. > for example, > {code} > $ yarn jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -shell_env YARN_CONTAINER_RUNTIME_TYPE=docker -shell_env > YARN_CONTAINER_RUNTIME_DOCKER_IMAGE=hadoop-docker -shell_command "date" -jar > $HADOOP_HOME/share/hadoop/yarn/hadoop-yarn-applications-distributedshell-2.7.1.jar > -num_containers 1 -timeout 360 > … > 17/04/12 00:16:40 INFO distributedshell.Client: Application did finished > unsuccessfully. YarnState=FINISHED, DSFinalStatus=FAILED. Breaking monitoring > loop > 17/04/12 00:16:40 ERROR distributedshell.Client: Application failed to > complete successfully > {code} > The error log is like below. > {code} > ... > Failed to write pid to file > /cgroup_parent/cpu/hadoop-yarn/container_/tasks - No such process > ... > {code} > When writing pid to cgroup tasks, container-executor doesn’t check docker > container’s status. > If the container finished very quickly, we can’t write pid to cgroup tasks, > and it is not problem. > So container-executor needs to check docker container’s exit code during > writing pid to cgroup tasks. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org