Hi Till, Thanks very much ! It's helpful.
HuangWHWHW 2017/1/16 -----邮件原件----- 发件人: Till Rohrmann [mailto:trohrm...@apache.org] 发送时间: 2017年1月15日 20:11 收件人: dev@flink.apache.org 抄送: u...@flink.apache.org 主题: Re: some questions about submit flink job on flink-yarn Hi Huang, the reason why you cannot use the IP address to send messages to your YARN JobManager is that we no longer resolve the hostname into an IP address. Instead we start the ActorSystem with the unresolved hostname. You can see this in the following log line: `Actor system bound to hostname 9-96-101-251`. Since Akka requires that the destination address of a message matches exactly the address to which the ActorSystem is bound, you have to use `9-96-101-251:38785`. This was recently changed. Concerning the ports, Flink chooses a random port for the `JobManager` in order to avoid port conflicts with other `JobManagers` running on the same node. With YARN you don't have control over where the `JobManager` is placed. However, you can use the configuration parameter `yarn.application-master.port` to specify a port or a port range for the application master/job manager. Additionally, the web frontends port is always overwritten and set to 0 which means random port selection when starting a yarn session. I hope this clarifies things a little bit. Cheers, Till On Sat, Jan 14, 2017 at 2:52 AM, huangwei (G) <huangwei...@huawei.com> wrote: > Hi Till, > > The "9-96-101-177" is just the hostname. > I rerun the flink on yarn and here is the jobmanager.log, and sorry > for I blocked some sensitive log. By the way , the port(another > questions in my earlier mail) seems to be a random value(this time is 38785 > and 35699). > I used flink-1.2.0, it works well on the yarn which is provided by > apache open source. But I run flink on a special yarn which was token > some safety reinforcement based on apache-yarn. I just have no idea about the > ERROR log. > > Jobmanage.log: > > 2017-01-14 09:24:35,584 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - > ------------------------------------------------------------ > -------------------- > 2017-01-14 09:24:35,585 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - Starting YARN ApplicationMaster / ResourceManager / > JobManager (Version: 1.2.0, Rev:82b1079, Date:04.01.2017 @ 17:38:23 > CST) > 2017-01-14 09:24:35,585 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - Current user: admin > 2017-01-14 09:24:35,585 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - JVM: Java HotSpot(TM) 64-Bit Server VM - Oracle > Corporation - > 1.8/25.112-b15 > 2017-01-14 09:24:35,585 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - Maximum heap size: 406 MiBytes > 2017-01-14 09:24:35,585 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - JAVA_HOME: /opt/huawei/Bigdata/jdk1.8.0_112/ > 2017-01-14 09:24:35,587 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - Hadoop version: 2.7.2 > 2017-01-14 09:24:35,587 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - JVM Options: > 2017-01-14 09:24:35,587 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - -Xmx424M > 2017-01-14 09:24:35,587 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - -Dlog.file=/srv/BigData/hadoop/data1/nm/containerlogs/ > application_1483499303549_0043/container_1483499303549_ > 0043_01_000001/jobmanager.log > 2017-01-14 09:24:35,587 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - -Dlogback.configurationFile=file:logback.xml > 2017-01-14 09:24:35,587 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - -Dlog4j.configuration=file:log4j.properties > 2017-01-14 09:24:35,587 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - Program Arguments: (none) > 2017-01-14 09:24:35,589 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - > ------------------------------------------------------------ > -------------------- > 2017-01-14 09:24:35,589 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - Registered UNIX signal handlers for [TERM, HUP, INT] > 2017-01-14 09:24:35,591 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - remoteKeytabPrincipal obtained admin > 2017-01-14 09:24:35,592 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - YARN daemon is running as: admin Yarn client user obtainer: > ad...@hadoop.com > 2017-01-14 09:24:35,596 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - Loading config from directory > /srv/BigData/hadoop/data1/nm/ > localdir/usercache/admin/appcache/application_ > 1483499303549_0043/container_1483499303549_0043_01_000001 > 2017-01-14 09:24:35,598 INFO > org.apache.flink.configuration.GlobalConfiguration > - Loading configuration property: jobmanager.rpc.address, > 9.96.101.32 > 2017-01-14 09:24:35,598 INFO > org.apache.flink.configuration.GlobalConfiguration > - Loading configuration property: jobmanager.rpc.port, 6123 > 2017-01-14 09:24:35,598 INFO > org.apache.flink.configuration.GlobalConfiguration > - Loading configuration property: jobmanager.heap.mb, 256 > 2017-01-14 09:24:35,598 INFO > org.apache.flink.configuration.GlobalConfiguration > - Loading configuration property: taskmanager.heap.mb, 512 > 2017-01-14 09:24:35,598 INFO > org.apache.flink.configuration.GlobalConfiguration > - Loading configuration property: > taskmanager.numberOfTaskSlots, 1 > 2017-01-14 09:24:35,598 INFO > org.apache.flink.configuration.GlobalConfiguration > - Loading configuration property: > taskmanager.memory.preallocate, false > 2017-01-14 09:24:35,599 INFO > org.apache.flink.configuration.GlobalConfiguration > - Loading configuration property: parallelism.default, 1 > 2017-01-14 09:24:35,599 INFO > org.apache.flink.configuration.GlobalConfiguration > - Loading configuration property: jobmanager.web.port, 8081 > 2017-01-14 09:24:35,599 INFO > org.apache.flink.configuration.GlobalConfiguration > - Loading configuration property: security.keytab, > /home/demo/flink/release/flink-1.2.0/keytab/user.keytab > 2017-01-14 09:24:35,599 INFO > org.apache.flink.configuration.GlobalConfiguration > - Loading configuration property: security.principal, admin > 2017-01-14 09:24:35,608 INFO > org.apache.flink.runtime.security.JaasConfiguration > - Initializing JAAS configuration instance. Parameters: > /srv/BigData/hadoop/data1/nm/localdir/usercache/admin/ > appcache/application_1483499303549_0043/container_ > 1483499303549_0043_01_000001/krb5.keytab, admin > 2017-01-14 09:24:35,609 INFO org.apache.flink.runtime.security.SecurityUtils > - SASL client auth for ZK will be disabled > 2017-01-14 09:24:35,824 INFO org.apache.hadoop.security.UserGroupInformation > - Login successful for user admin using keytab file > /srv/BigData/hadoop/data1/nm/localdir/usercache/admin/ > appcache/application_1483499303549_0043/container_ > 1483499303549_0043_01_000001/krb5.keytab > 2017-01-14 09:24:35,825 INFO org.apache.flink.runtime.security.SecurityUtils > - Hadoop user set to ad...@hadoop.com (auth:KERBEROS) > 2017-01-14 09:24:35,936 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - YARN assigned hostname for application master: > 9-96-101-251 > 2017-01-14 09:24:35,936 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - keytabPath: /srv/BigData/hadoop/data1/nm/ > localdir/usercache/admin/appcache/application_ > 1483499303549_0043/container_1483499303549_0043_01_000001/krb5.keytab > 2017-01-14 09:24:35,938 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - TaskManagers will be created with 1 task slots > 2017-01-14 09:24:35,938 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - TaskManagers will be started with container size 1024 MB, > JVM heap size 424 MB, JVM direct memory limit 424 MB > 2017-01-14 09:24:35,943 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - Trying to start actor system at 9.96.101.251:38785 > 2017-01-14 09:24:36,336 INFO akka.event.slf4j.Slf4jLogger > - Slf4jLogger started > 2017-01-14 09:24:36,438 INFO Remoting > - Starting remoting > 2017-01-14 09:24:36,547 INFO Remoting > - Remoting started; listening on addresses > :[akka.tcp://flink@9-96-101-251:38785] > 2017-01-14 09:24:36,551 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - Actor system started at 9.96.101.251:38785 > 2017-01-14 09:24:36,551 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - Actor system bound to hostname 9-96-101-251. > 2017-01-14 09:24:36,554 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - Setting up resources for TaskManagers > 2017-01-14 09:24:36,554 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - TM:remoteKeytabPath obtained hdfs://hacluster/user/admin/. > flink/application_1483499303549_0043/user.keytab > 2017-01-14 09:24:36,555 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - TM:remoteKeytabPrincipal obtained admin > 2017-01-14 09:24:36,555 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - TM:remoteYarnConfPath obtained null > 2017-01-14 09:24:36,555 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - TM:remotekrb5Path obtained null > 2017-01-14 09:24:36,932 WARN org.apache.hadoop.util.NativeCodeLoader > - Unable to load native-hadoop library for your > platform... using builtin-java classes where applicable > 2017-01-14 09:24:36,945 WARN > org.apache.hadoop.hdfs.shortcircuit.DomainSocketFactory > - The short-circuit local reads feature cannot be used because > libhadoop cannot be loaded. > 2017-01-14 09:24:36,949 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - Adding keytab > hdfs://hacluster/user/admin/.flink/application_1483499303549_0043/user > .keytab to the AM container local resource bucket > 2017-01-14 09:24:37,085 INFO org.apache.flink.yarn.Utils > - Copying from file:/srv/BigData/hadoop/ > data1/nm/localdir/usercache/admin/appcache/application_ > 1483499303549_0043/container_1483499303549_0043_01_000001/ > e2943789-d80a-4abd-8ae9-2fc14cb1fd03-taskmanager-conf.yaml to > hdfs://hacluster/user/admin/.flink/application_ > 1483499303549_0043/e2943789-d80a-4abd-8ae9-2fc14cb1fd03- > taskmanager-conf.yaml > 2017-01-14 09:24:37,258 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - Prepared local resource for modified yaml: resource { scheme: > "hdfs" host: "hacluster" port: -1 file: "/user/admin/.flink/ > application_1483499303549_0043/e2943789-d80a-4abd-8ae9- > 2fc14cb1fd03-taskmanager-conf.yaml" } size: 878 timestamp: > 1484357077250 > type: FILE visibility: APPLICATION > 2017-01-14 09:24:37,265 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - Creating container launch context for TaskManagers > 2017-01-14 09:24:37,265 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - Starting TaskManagers with command: $JAVA_HOME/bin/java > -Xms424m -Xmx424m -XX:MaxDirectMemorySize=424m > -Dlog.file=<LOG_DIR>/taskmanager.log > -Dlogback.configurationFile=file:./logback.xml > -Dlog4j.configuration=file:./log4j.properties > org.apache.flink.yarn.YarnTaskManager > --configDir . 1> <LOG_DIR>/taskmanager.out 2> > <LOG_DIR>/taskmanager.err > 2017-01-14 09:24:37,288 INFO org.apache.flink.runtime.blob.BlobServer > - Created BLOB server storage directory > /tmp/blobStore-f91bad88-1473-4e86-b151-f93dffa58baa > 2017-01-14 09:24:37,289 INFO org.apache.flink.runtime.blob.BlobServer > - Started BLOB server at 0.0.0.0:53972 - max > concurrent > requests: 50 - max backlog: 1000 > 2017-01-14 09:24:37,302 INFO org.apache.flink.runtime.metrics.MetricRegistry > - No metrics reporter configured, no metrics will be > exposed/reported. > 2017-01-14 09:24:37,307 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - Starting JobManager Web Frontend > 2017-01-14 09:24:37,310 INFO > org.apache.flink.runtime.jobmanager.MemoryArchivist > - Started memory archivist akka://flink/user/$a > 2017-01-14 09:24:37,311 INFO org.apache.flink.yarn.YarnJobManager > - Starting JobManager at akka.tcp://flink@9-96-101-251: > 38785/user/jobmanager. > 2017-01-14 09:24:37,318 INFO > org.apache.flink.runtime.webmonitor.WebMonitorUtils > - Determined location of JobManager log file: > /srv/BigData/hadoop/data1/nm/containerlogs/application_ > 1483499303549_0043/container_1483499303549_0043_01_000001/jobmanager.l > og > 2017-01-14 09:24:37,318 INFO > org.apache.flink.runtime.webmonitor.WebMonitorUtils > - Determined location of JobManager stdout file: > /srv/BigData/hadoop/data1/nm/containerlogs/application_ > 1483499303549_0043/container_1483499303549_0043_01_000001/jobmanager.o > ut > 2017-01-14 09:24:37,318 INFO > org.apache.flink.runtime.webmonitor.WebRuntimeMonitor > - Using directory > /tmp/flink-web-c4991b46-e637-4207-80ed-caef4cf5702e > for the web interface files > 2017-01-14 09:24:37,364 INFO > org.apache.flink.runtime.webmonitor.WebRuntimeMonitor > - Using directory > /tmp/flink-web-f8b7b6d7-b066-48a1-9536-40a3cdf42778 > for web frontend JAR file uploads > 2017-01-14 09:24:37,378 INFO org.apache.flink.yarn.YarnJobManager > - JobManager > akka.tcp://flink@9-96-101-251:38785/user/jobmanager > was granted leadership with leader session ID None. > 2017-01-14 09:24:37,560 INFO > org.apache.flink.runtime.webmonitor.WebRuntimeMonitor > - Web frontend listening at 0:0:0:0:0:0:0:0:35699 > 2017-01-14 09:24:37,561 INFO > org.apache.flink.runtime.webmonitor.WebRuntimeMonitor > - Starting with JobManager > akka.tcp://flink@9-96-101-251:38785/user/jobmanager > on port 35699 > 2017-01-14 09:24:37,561 INFO > org.apache.flink.runtime.webmonitor.JobManagerRetriever > - New leader reachable under akka://flink/user/jobmanager#- > 640052308:null. > 2017-01-14 09:24:37,568 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - YARN application tolerates 4 failed TaskManager > containers before giving up > 2017-01-14 09:24:37,571 INFO > org.apache.flink.yarn.YarnApplicationMasterRunner > - YARN Application Master started > 2017-01-14 09:24:37,579 INFO org.apache.flink.yarn.YarnFlinkResourceManager > - Initializing YARN resource master > 2017-01-14 09:24:37,605 INFO org.apache.hadoop.yarn.client.api.impl. > ContainerManagementProtocolProxy - > yarn.client.max-cached-nodemanagers-proxies > : 0 > 2017-01-14 09:24:37,606 INFO org.apache.flink.yarn.YarnFlinkResourceManager > - Registering Application Master with tracking url > http://9-96-101-251:35699 > 2017-01-14 09:24:37,641 INFO org.apache.flink.yarn.YarnFlinkResourceManager > - Trying to associate with JobManager leader > akka://flink/user/jobmanager#-640052308 > 2017-01-14 09:24:37,647 INFO org.apache.flink.yarn.YarnFlinkResourceManager > - Resource Manager associating with leading JobManager > Actor[akka://flink/user/jobmanager#-640052308] - leader session null > 2017-01-14 09:24:37,648 INFO org.apache.flink.yarn.YarnFlinkResourceManager > - Requesting new TaskManager container with 1024 > megabytes memory. Pending requests: 1 > 2017-01-14 09:24:37,654 INFO org.apache.flink.yarn.YarnFlinkResourceManager > - Requesting new TaskManager container with 1024 > megabytes memory. Pending requests: 2 > 2017-01-14 09:24:37,654 INFO org.apache.flink.yarn.YarnFlinkResourceManager > - Requesting new TaskManager container with 1024 > megabytes memory. Pending requests: 3 > 2017-01-14 09:24:37,655 INFO org.apache.flink.yarn.YarnFlinkResourceManager > - Requesting new TaskManager container with 1024 > megabytes memory. Pending requests: 4 > 2017-01-14 09:24:38,681 INFO > org.apache.hadoop.yarn.client.api.impl.AMRMClientImpl > - Received new token for : 9-96-101-177:26009 > 2017-01-14 09:24:38,692 INFO org.apache.flink.yarn.YarnFlinkResourceManager > - Received new container: > container_1483499303549_0043_01_000002 > - Remaining pending container requests: 3 > 2017-01-14 09:24:38,693 INFO org.apache.flink.yarn.YarnFlinkResourceManager > - Launching TaskManager in container ContainerInLaunch @ > 1484357078692: Container: [ContainerId: > container_1483499303549_0043_01_000002, > NodeId: 9-96-101-177:26009, NodeHttpAddress: 9-96-101-177:26010, Resource: > <memory:1024, vCores:1>, Priority: 0, Token: Token { kind: > ContainerToken, > service: 9.96.101.177:26009 }, ] on host 9-96-101-177 > 2017-01-14 09:24:38,694 INFO org.apache.hadoop.yarn.client.api.impl. > ContainerManagementProtocolProxy - Opening proxy : 9-96-101-177:26009 > 2017-01-14 09:24:39,189 INFO > org.apache.hadoop.yarn.client.api.impl.AMRMClientImpl > - Received new token for : 9-96-101-251:26009 > 2017-01-14 09:24:39,189 INFO > org.apache.hadoop.yarn.client.api.impl.AMRMClientImpl > - Received new token for : 9-96-101-32:26009 > 2017-01-14 09:24:39,189 INFO > org.apache.hadoop.yarn.client.api.impl.AMRMClientImpl > - Received new token for : 9-91-8-160:26009 > 2017-01-14 09:24:39,190 INFO org.apache.flink.yarn.YarnFlinkResourceManager > - Received new container: > container_1483499303549_0043_01_000003 > - Remaining pending container requests: 2 > 2017-01-14 09:24:39,190 INFO org.apache.flink.yarn.YarnFlinkResourceManager > - Launching TaskManager in container ContainerInLaunch @ > 1484357079190: Container: [ContainerId: > container_1483499303549_0043_01_000003, > NodeId: 9-96-101-251:26009, NodeHttpAddress: 9-96-101-251:26010, Resource: > <memory:1024, vCores:1>, Priority: 0, Token: Token { kind: > ContainerToken, > service: 9.96.101.251:26009 }, ] on host 9-96-101-251 > 2017-01-14 09:24:39,190 INFO org.apache.hadoop.yarn.client.api.impl. > ContainerManagementProtocolProxy - Opening proxy : 9-96-101-251:26009 > 2017-01-14 09:24:39,202 INFO org.apache.flink.yarn.YarnFlinkResourceManager > - Received new container: > container_1483499303549_0043_01_000004 > - Remaining pending container requests: 1 > 2017-01-14 09:24:39,202 INFO org.apache.flink.yarn.YarnFlinkResourceManager > - Launching TaskManager in container ContainerInLaunch @ > 1484357079202: Container: [ContainerId: > container_1483499303549_0043_01_000004, > NodeId: 9-96-101-32:26009, NodeHttpAddress: 9-96-101-32:26010, Resource: > <memory:1024, vCores:1>, Priority: 0, Token: Token { kind: > ContainerToken, > service: 9.96.101.32:26009 }, ] on host 9-96-101-32 > 2017-01-14 09:24:39,202 INFO org.apache.hadoop.yarn.client.api.impl. > ContainerManagementProtocolProxy - Opening proxy : 9-96-101-32:26009 > 2017-01-14 09:24:39,217 INFO org.apache.flink.yarn.YarnFlinkResourceManager > - Received new container: > container_1483499303549_0043_01_000005 > - Remaining pending container requests: 0 > 2017-01-14 09:24:39,217 INFO org.apache.flink.yarn.YarnFlinkResourceManager > - Launching TaskManager in container ContainerInLaunch @ > 1484357079217: Container: [ContainerId: > container_1483499303549_0043_01_000005, > NodeId: 9-91-8-160:26009, NodeHttpAddress: 9-91-8-160:26010, Resource: > <memory:1024, vCores:1>, Priority: 0, Token: Token { kind: > ContainerToken, > service: 9.91.8.160:26009 }, ] on host 9-91-8-160 > 2017-01-14 09:24:39,217 INFO org.apache.hadoop.yarn.client.api.impl. > ContainerManagementProtocolProxy - Opening proxy : 9-91-8-160:26009 > 2017-01-14 09:24:43,348 INFO org.apache.flink.yarn.YarnFlinkResourceManager > - TaskManager container_1483499303549_0043_01_000003 has > started. > 2017-01-14 09:24:43,350 INFO > org.apache.flink.runtime.instance.InstanceManager > - Registered TaskManager at 9-96-101-251 > (akka.tcp://flink@9-96-101-251:57010/user/taskmanager) as > 49800ab8cfcd1a11e45084a48281df75. Current number of registered hosts > is 1. Current number of alive task slots is 1. > 2017-01-14 09:24:44,881 INFO org.apache.flink.yarn.YarnFlinkResourceManager > - TaskManager container_1483499303549_0043_01_000002 has > started. > 2017-01-14 09:24:44,881 INFO > org.apache.flink.runtime.instance.InstanceManager > - Registered TaskManager at 9-96-101-177 > (akka.tcp://flink@9-96-101-177:35778/user/taskmanager) as > 5d65baf1ec196cf3ac5bc43870156855. Current number of registered hosts > is 2. Current number of alive task slots is 2. > 2017-01-14 09:24:45,855 INFO org.apache.flink.yarn.YarnFlinkResourceManager > - TaskManager container_1483499303549_0043_01_000004 has > started. > 2017-01-14 09:24:45,855 INFO > org.apache.flink.runtime.instance.InstanceManager > - Registered TaskManager at 9-96-101-32 > (akka.tcp://flink@9-96-101-32:58486/user/taskmanager) as > 80de4a47fa60536b78ea052cbadec7ee. Current number of registered hosts > is 3. Current number of alive task slots is 3. > 2017-01-14 09:24:46,018 INFO org.apache.flink.yarn.YarnFlinkResourceManager > - TaskManager container_1483499303549_0043_01_000005 has > started. > 2017-01-14 09:24:46,018 INFO > org.apache.flink.runtime.instance.InstanceManager > - Registered TaskManager at 9-91-8-160 > (akka.tcp://flink@9-91-8-160:47548/user/taskmanager) as > e9f2497d6223b2d704b3aced665a3c02. Current number of registered hosts > is 4. Current number of alive task slots is 4. > 2017-01-14 09:29:58,066 ERROR akka.remote.EndpointWriter > - dropping message [class > akka.actor.ActorSelectionMessage] for non-local recipient > [Actor[akka.tcp://flink@9.96.101.251:38785/]] > arriving at [akka.tcp://flink@9.96.101.251:38785] inbound addresses > are [akka.tcp://flink@9-96-101-251:38785] > > Thanks! > > HuangWHWHW > 2017/1/14 > > -----邮件原件----- > 发件人: Till Rohrmann [mailto:trohrm...@apache.org] > 发送时间: 2017年1月13日 18:22 > 收件人: dev@flink.apache.org > 抄送: u...@flink.apache.org > 主题: Re: some questions about submit flink job on flink-yarn > > Hi Huang, > > this seems to be very strange, because the JobManager’s actor system > has bound to the address 9-96-101-177 instead of 9.96.101.177. It > seems a if the . have been replaced by -. > > Could you maybe tell me which version of Flink you’re running and also > share the complete JobManager log with us? > > I tested it with the latest 1.2 SNAPSHOT version and there it seemed > to work. > > Cheers, > Till > > > On Fri, Jan 13, 2017 at 9:02 AM, huangwei (G) <huangwei...@huawei.com> > wrote: > > > Dear All, > > > > I get an error in jobmanage.log following when I submit a flink job > > (batch/WordCount.jar) by using command : "./bin/flink run -m > > 9.96.101.177:39180 ./examples/batch/WordCount.jar". > > > > And the flink is on yarn cluster. > > > > Error in jobmanage.log : > > 2017-01-13 15:28:27,402 ERROR akka.remote.EndpointWriter > > - dropping message [class > > akka.actor.ActorSelectionMessage] for non-local recipient > > [Actor[akka.tcp://flink@9.96.101.177:39180/]] > > arriving at [akka.tcp://flink@9.96.101.177:39180] inbound addresses > > are [akka.tcp://flink@9-96-101-177:39180] > > > > However, It is success when I use flink web-ui to submit the job. > > > > How to solve this problem? > > > > And otherwise, when I started the flink on yarn, the > > jobmanage.rpc.port and the web port both were changed to 39180 and 57724. > > The configuration following in flink-conf.yaml is just as default : > > > > jobmanager.rpc.port: 6123 > > > > and > > > > jobmanager.web.port: 8081 > > > > I started the flink on yarn using command : "./bin/yarn-session.sh -n 4". > > > > Why were the ports changed to 39180 and 57724? > > > > Many thanks if there is any help! > > > > HuangWHWHW > > 2017.1.13 > > >
