Hi, 从异常日志看,应该是因为你的state.checkpoints.dir 或者说 statebackend的checkpoint目录配置成了本地目录,checkpoint保存到了本地机器上,所以在failover restore的时候,必须得让原task部署回原来的机器才能正常运行。将state backend的checkpoint目录更换为一个DFS目录即可。
祝好 唐云 ________________________________ From: Carmen Free <ddjxuxi...@gmail.com> Sent: Tuesday, January 12, 2021 18:14 To: user-zh@flink.apache.org <user-zh@flink.apache.org> Subject: Re: rocksdb作为statebackend时,TM节点挂掉了,为何任务不能恢复呢? 你好,唐老师,谢谢解答。 不好意思,下面补充一下报错信息,刚才忘记说了。 主要报错信息如下,重新模拟了下: 2021-01-12 18:09:34,236 INFO org.apache.flink.runtime.executiongraph.ExecutionGraph - Source: Custom Source -> Flat Map -> Timestamps/Watermarks (2/2) (3b50a7ce56b408c2978260846b76a28a) switched from RUNNING to FAILED on org.apache.flink.runtime.jobmaster.slotpool.SingleLogicalSlot@4c107a3a. java.lang.Exception: Exception while creating StreamOperatorStateContext. at org.apache.flink.streaming.api.operators.StreamTaskStateInitializerImpl.streamOperatorStateContext(StreamTaskStateInitializerImpl.java:191) at org.apache.flink.streaming.api.operators.AbstractStreamOperator.initializeState(AbstractStreamOperator.java:255) at org.apache.flink.streaming.runtime.tasks.StreamTask.initializeStateAndOpen(StreamTask.java:989) at org.apache.flink.streaming.runtime.tasks.StreamTask.lambda$beforeInvoke$0(StreamTask.java:453) at org.apache.flink.streaming.runtime.tasks.StreamTaskActionExecutor$SynchronizedStreamTaskActionExecutor.runThrowing(StreamTaskActionExecutor.java:94) at org.apache.flink.streaming.runtime.tasks.StreamTask.beforeInvoke(StreamTask.java:448) at org.apache.flink.streaming.runtime.tasks.StreamTask.invoke(StreamTask.java:460) at org.apache.flink.runtime.taskmanager.Task.doRun(Task.java:708) at org.apache.flink.runtime.taskmanager.Task.run(Task.java:533) at java.lang.Thread.run(Thread.java:748) Caused by: org.apache.flink.util.FlinkException: Could not restore operator state backend for StreamSource_cbc357ccb763df2852fee8c4fc7d55f2_(2/2) from any of the 1 provided restore options. at org.apache.flink.streaming.api.operators.BackendRestorerProcedure.createAndRestore(BackendRestorerProcedure.java:135) at org.apache.flink.streaming.api.operators.StreamTaskStateInitializerImpl.operatorStateBackend(StreamTaskStateInitializerImpl.java:252) at org.apache.flink.streaming.api.operators.StreamTaskStateInitializerImpl.streamOperatorStateContext(StreamTaskStateInitializerImpl.java:139) ... 9 more Caused by: org.apache.flink.runtime.state.BackendBuildingException: Failed when trying to restore operator state backend at org.apache.flink.runtime.state.DefaultOperatorStateBackendBuilder.build(DefaultOperatorStateBackendBuilder.java:86) at org.apache.flink.contrib.streaming.state.RocksDBStateBackend.createOperatorStateBackend(RocksDBStateBackend.java:565) at org.apache.flink.streaming.api.operators.StreamTaskStateInitializerImpl.lambda$operatorStateBackend$0(StreamTaskStateInitializerImpl.java:243) at org.apache.flink.streaming.api.operators.BackendRestorerProcedure.attemptCreateAndRestore(BackendRestorerProcedure.java:142) at org.apache.flink.streaming.api.operators.BackendRestorerProcedure.createAndRestore(BackendRestorerProcedure.java:121) ... 11 more Caused by: java.io.FileNotFoundException: /data/flink/checkpoints/b261de0447d59bb2f1db9c084f5b1a0b/chk-5/4a160901-8ddd-468e-a82d-6efcb8a9dff9 (No such file or directory) at java.io.FileInputStream.open0(Native Method) at java.io.FileInputStream.open(FileInputStream.java:195) at java.io.FileInputStream.<init>(FileInputStream.java:138) at org.apache.flink.core.fs.local.LocalDataInputStream.<init>(LocalDataInputStream.java:50) at org.apache.flink.core.fs.local.LocalFileSystem.open(LocalFileSystem.java:142) at org.apache.flink.core.fs.SafetyNetWrapperFileSystem.open(SafetyNetWrapperFileSystem.java:85) at org.apache.flink.runtime.state.filesystem.FileStateHandle.openInputStream(FileStateHandle.java:68) at org.apache.flink.runtime.state.OperatorStreamStateHandle.openInputStream(OperatorStreamStateHandle.java:66) at org.apache.flink.runtime.state.OperatorStateRestoreOperation.restore(OperatorStateRestoreOperation.java:73) at org.apache.flink.runtime.state.DefaultOperatorStateBackendBuilder.build(DefaultOperatorStateBackendBuilder.java:83) ... 15 more 这个文件夹在A节点(JM)上是有的,难道是访问权限问题吗?B节点无法访问A节点吗,有点奇怪啊,配置了ssh免密的啊,文件夹/data/flink/checkpoints访问权限也设置成了777 Yun Tang <myas...@live.com> 于2021年1月12日周二 下午5:46写道: > Hi > > Flink的容错机制是可以保证TM lost时候会尝试重启作业,“为何任务不能恢复”是需要看完整异常栈的,简单描述是无法帮助排查问题的。 > > 祝好 > 唐云 > ________________________________ > From: Carmen Free <ddjxuxi...@gmail.com> > Sent: Tuesday, January 12, 2021 15:52 > To: user-zh@flink.apache.org <user-zh@flink.apache.org> > Subject: rocksdb作为statebackend时,TM节点挂掉了,为何任务不能恢复呢? > > hi, > > rocksdb作为statebackend时,TM节点挂掉了,为何任务不能恢复呢? > > 1、环境说明 > > flink版本:1.10.2 > 操作系统:centos 7 > > 2、集群说明(当前模拟了2节点) > > 节点A | 节点B > 角色 | JM、TM | TM > taskslot | 4 | 4 > > 3、statebackend配置 > > # rocksdb作为状态后备 > state.backend: rocksdb > > > # 存储快照的目录(暂时使用的本地目录) > state.checkpoints.dir: file:///data/flink/checkpoints > > > 4、启动任务后,任务自动分配在A节点的TM上,运行一段时间后,检查点快照正常。接着,仅停掉A节点TM(JM仍正常运行),任务被自动调度至B节点的TM上,但是此时任务一直重启,无法恢复,这是为什么呢? > > > 5、如果我启动A节点,任务依旧无法恢复(此时任务仍在B节点运行),直到我停掉B节点TM,此时任务调度至A节点,任务可以正常恢复。所以有点疑问,4中的场景为何不能恢复任务呢?为什么只有在A节点上才可以进行任务恢复呢?最初以为是访问路径的问题,但是仔细想了想,检查点相关的操作一直都是JM进行的,我觉得只要JM没有挂掉,应该就可以将任务进行恢复啊,是我的理解有偏差吗? >