I'm not aware of any significant changes to the HA components between 1.9/1.11. Would you mind sharing the complete jobmanager/taskmanager logs?
Thank you~ Xintong Song On Fri, Dec 18, 2020 at 8:53 AM Lu Niu <qqib...@gmail.com> wrote: > Hi, Xintong > > Thanks for replying and your suggestion. I did check the ZK side but there > is nothing interesting. The error message actually shows that only one TM > thought JM lost leadership while others ran fine. Also, this happened only > after we migrated from 1.9 to 1.11. Is it possible this is introduced by > 1.11? > > Best > Lu > > On Wed, Dec 16, 2020 at 5:56 PM Xintong Song <tonysong...@gmail.com> > wrote: > >> Hi Lu, >> >> I assume you are using ZooKeeper as the HA service? >> >> A common cause of unexpected leadership lost is the instability of HA >> service. E.g., if ZK does not receive heartbeat from Flink RM for a >> certain period of time, it will revoke the leadership and notify >> other components. You can look into the ZooKeeper logs checking why RM's >> leadership is revoked. >> >> Thank you~ >> >> Xintong Song >> >> >> >> On Thu, Dec 17, 2020 at 8:42 AM Lu Niu <qqib...@gmail.com> wrote: >> >>> Hi, Flink users >>> >>> Recently we migrated to flink 1.11 and see exceptions like: >>> ``` >>> 2020-12-15 12:41:01,199 INFO >>> org.apache.flink.runtime.executiongraph.ExecutionGraph [] - Source: >>> USER_MATERIALIZED_EVENT_SIGNAL-user_context-event -> >>> USER_MATERIALIZED_EVENT_SIGNAL-user_context-event-as_nrtgtuple (21/60) >>> (711d1d319691a4b80e30fe6ab7dfab5b) switched from RUNNING to FAILED on >>> org.apache.flink.runtime.jobmaster.slotpool.SingleLogicalSlot@50abf386. >>> java.lang.Exception: Job leader for job id >>> 47b1531f79ffe3b86bc5910f6071e40c lost leadership. >>> at >>> org.apache.flink.runtime.taskexecutor.TaskExecutor$JobLeaderListenerImpl.lambda$null$2(TaskExecutor.java:1852) >>> ~[nrtg-1.11_deploy.jar:?] >>> at java.util.Optional.ifPresent(Optional.java:159) ~[?:1.8.0_212-ga] >>> at >>> org.apache.flink.runtime.taskexecutor.TaskExecutor$JobLeaderListenerImpl.lambda$jobManagerLostLeadership$3(TaskExecutor.java:1851) >>> ~[nrtg-1.11_deploy.jar:?] >>> at >>> org.apache.flink.runtime.rpc.akka.AkkaRpcActor.handleRunAsync(AkkaRpcActor.java:402) >>> ~[nrtg-1.11_deploy.jar:?] >>> at >>> org.apache.flink.runtime.rpc.akka.AkkaRpcActor.handleRpcMessage(AkkaRpcActor.java:195) >>> ~[nrtg-1.11_deploy.jar:?] >>> at >>> org.apache.flink.runtime.rpc.akka.AkkaRpcActor.handleMessage(AkkaRpcActor.java:152) >>> ~[nrtg-1.11_deploy.jar:?] >>> at akka.japi.pf.UnitCaseStatement.apply(CaseStatements.scala:24) >>> [nrtg-1.11_deploy.jar:?] >>> at akka.japi.pf.UnitCaseStatement.apply(CaseStatements.scala:20) >>> [nrtg-1.11_deploy.jar:?] >>> at scala.PartialFunction$class.applyOrElse(PartialFunction.scala:123) >>> [nrtg-1.11_deploy.jar:?] >>> at akka.japi.pf.UnitCaseStatement.applyOrElse(CaseStatements.scala:20) >>> [nrtg-1.11_deploy.jar:?] >>> at scala.PartialFunction$OrElse.applyOrElse(PartialFunction.scala:170) >>> [nrtg-1.11_deploy.jar:?] >>> at scala.PartialFunction$OrElse.applyOrElse(PartialFunction.scala:171) >>> [nrtg-1.11_deploy.jar:?] >>> at scala.PartialFunction$OrElse.applyOrElse(PartialFunction.scala:171) >>> [nrtg-1.11_deploy.jar:?] >>> at akka.actor.Actor$class.aroundReceive(Actor.scala:539) >>> [nrtg-1.11_deploy.jar:?] >>> at akka.actor.AbstractActor.aroundReceive(AbstractActor.scala:227) >>> [nrtg-1.11_deploy.jar:?] >>> at akka.actor.ActorCell.receiveMessage(ActorCell.scala:612) >>> [nrtg-1.11_deploy.jar:?] >>> at akka.actor.ActorCell.invoke(ActorCell.scala:581) >>> [nrtg-1.11_deploy.jar:?] >>> at akka.dispatch.Mailbox.processMailbox(Mailbox.scala:268) >>> [nrtg-1.11_deploy.jar:?] >>> at akka.dispatch.Mailbox.run(Mailbox.scala:229) [nrtg-1.11_deploy.jar:?] >>> ``` >>> >>> ``` >>> 2020-12-15 01:01:39,531 INFO >>> org.apache.flink.runtime.executiongraph.ExecutionGraph [] - >>> USER_MATERIALIZED_EVENT_SIGNAL.user_context.SINK-stream_joiner -> >>> USER_MATERIALIZED_EVENT_SIGNAL-user_context-SINK-SINKS.realpin (260/360) >>> (0c1f4495088ec9452c597f46a88a2c8e) switched from RUNNING to FAILED on >>> org.apache.flink.runtime.jobmaster.slotpool.SingleLogicalSlot@2362b2fd. >>> org.apache.flink.util.FlinkException: ResourceManager leader changed to >>> new address null >>> at >>> org.apache.flink.runtime.taskexecutor.TaskExecutor.notifyOfNewResourceManagerLeader(TaskExecutor.java:1093) >>> ~[nrtg-1.11_deploy.jar:?] >>> at >>> org.apache.flink.runtime.taskexecutor.TaskExecutor.access$800(TaskExecutor.java:173) >>> ~[nrtg-1.11_deploy.jar:?] >>> at >>> org.apache.flink.runtime.taskexecutor.TaskExecutor$ResourceManagerLeaderListener.lambda$notifyLeaderAddress$0(TaskExecutor.java:1816) >>> ~[nrtg-1.11_deploy.jar:?] >>> at >>> org.apache.flink.runtime.rpc.akka.AkkaRpcActor.handleRunAsync(AkkaRpcActor.java:402) >>> ~[nrtg-1.11_deploy.jar:?] >>> at >>> org.apache.flink.runtime.rpc.akka.AkkaRpcActor.handleRpcMessage(AkkaRpcActor.java:195) >>> ~[nrtg-1.11_deploy.jar:?] >>> at >>> org.apache.flink.runtime.rpc.akka.AkkaRpcActor.handleMessage(AkkaRpcActor.java:152) >>> ~[nrtg-1.11_deploy.jar:?] >>> at akka.japi.pf.UnitCaseStatement.apply(CaseStatements.scala:24) >>> [nrtg-1.11_deploy.jar:?] >>> at akka.japi.pf.UnitCaseStatement.apply(CaseStatements.scala:20) >>> [nrtg-1.11_deploy.jar:?] >>> at scala.PartialFunction$class.applyOrElse(PartialFunction.scala:123) >>> [nrtg-1.11_deploy.jar:?] >>> at akka.japi.pf.UnitCaseStatement.applyOrElse(CaseStatements.scala:20) >>> [nrtg-1.11_deploy.jar:?] >>> at scala.PartialFunction$OrElse.applyOrElse(PartialFunction.scala:170) >>> [nrtg-1.11_deploy.jar:?] >>> at scala.PartialFunction$OrElse.applyOrElse(PartialFunction.scala:171) >>> [nrtg-1.11_deploy.jar:?] >>> at scala.PartialFunction$OrElse.applyOrElse(PartialFunction.scala:171) >>> [nrtg-1.11_deploy.jar:?] >>> at akka.actor.Actor$class.aroundReceive(Actor.scala:539) >>> [nrtg-1.11_deploy.jar:?] >>> at akka.actor.AbstractActor.aroundReceive(AbstractActor.scala:227) >>> [nrtg-1.11_deploy.jar:?] >>> at akka.actor.ActorCell.receiveMessage(ActorCell.scala:612) >>> [nrtg-1.11_deploy.jar:?] >>> at akka.actor.ActorCell.invoke(ActorCell.scala:581) >>> [nrtg-1.11_deploy.jar:?] >>> at akka.dispatch.Mailbox.processMailbox(Mailbox.scala:268) >>> [nrtg-1.11_deploy.jar:?] >>> at akka.dispatch.Mailbox.run(Mailbox.scala:229) [nrtg-1.11_deploy.jar:?] >>> ``` >>> >>> This happens a few times per week. It seems like one Task Manager >>> wrongly thought JobMananger is lost and triggers a full restart of the >>> whole job. Does anyone know how to resolve such errors? Thanks! >>> >>> Best >>> Lu >>> >>