> > 偶尔会出现checkpint需要超过30分钟的（出现的频率不高，1～2天1次）
找到原因了， 任务处理延迟比较大，kafka数据过期清理了，导致从last消费（watermark一下子增长了好多个小时）， 然后，这个时候需要输出几个小时内的所有窗口（平时一次只输出一个窗口，这时一次要输出30个窗口消耗比较长时间）。 因为是稳定测试任务，没有关注kafka 延迟 导致数据过期到问题。 感谢，zhisheng、LakeShen、Yichao Yang。 Yichao Yang <[email protected]> 于2020年6月29日周一 下午7:58写道： > Hi > > > > 如果任务会经常出现反压的话，可以先解决反压问题，因为任务反压也会影响checkpoint，其次就是可以关注一下你的作业的物理资源指标，比如cpu使用率，内存使用率，gc频率是否特别高，尝试保障物理资源使用率在正常水平。 > > > Best, > Yichao Yang > > > > > ------------------ 原始邮件 ------------------ > 发件人: "Tianwang Li"<[email protected]>; > 发送时间: 2020年6月28日(星期天) 上午10:17 > 收件人: "user-zh"<[email protected]>; > > 主题: Flink-1.10.0 source的checkpoint偶尔时间比较长 > > > > 关于Flink checkpoint偶尔会比较长时间的问题。 > 环境与背景： > 版本：flink1.10.0 > 数据量：每秒约10万左右的记录，数据源是kafka > 计算逻辑：滑动窗口统计，每个窗口输出的规模大概1～2千万记录。 > 是否有反压：任务经常会出现反压（特别是在窗口输出的时候）。 > > > 问题： > > 大部分时候checkpoint都是在1分钟内完成，偶尔会出现checkpint需要超过30分钟的（出现的频率不高，1～2天1次）。 > source的checkpoint消耗的时间比较长。Trigger checkpoint 到 Starting > checkpoint消耗时间比较长。 > > > checkpoint情况大致如下： > > > > > > > > > > > > > > > 2020-06-24 21:09:53,369 DEBUG > org.apache.flink.runtime.taskexecutor.TaskExecutor > - Trigger checkpoint 316@1593004193363 for > 84dce1ec8aa5a4df2d1758d6e9278693. > > 2020-06-24 21:09:58,327 DEBUG > org.apache.flink.runtime.taskexecutor.TaskExecutor > - Received heartbeat request from > e88ea2f790430c9c160e540ef0546d60. > > 2020-06-24 21:09:59,266 DEBUG > org.apache.flink.runtime.taskexecutor.TaskExecutor > - Received heartbeat request from > b93d7167db364dfdcbda886944f1482f. > > 2020-06-24 21:09:59,686 INFO > org.apache.flink.runtime.taskexecutor.TaskManagerRunner > - Memory usage stats: [HEAP: 202/2573/2573 MB, NON HEAP: 111/114/424 > MB (used/committed/max)] > > 2020-06-24 21:09:59,686 INFO > org.apache.flink.runtime.taskexecutor.TaskManagerRunner > - Direct memory stats: Count: 17403, Total Capacity: 583911423, Used > Memory: 583911424 > > 2020-06-24 21:09:59,686 INFO > org.apache.flink.runtime.taskexecutor.TaskManagerRunner > - Off-heap pool stats: [Code Cache: 35/35/240 MB > (used/committed/max)], [Metaspace: 67/69/96 MB (used/committed/max)], > [Compressed Class Space: 8/9/88 MB (used/committed/max)] > > 2020-06-24 21:09:59,686 INFO > org.apache.flink.runtime.taskexecutor.TaskManagerRunner > - Garbage collector stats: [PS Scavenge, GC TIME (ms): 108643, GC > COUNT: 6981], [PS MarkSweep, GC TIME (ms): 1074, GC COUNT: 6] > > 2020-06-24 21:10:08,346 DEBUG > org.apache.flink.runtime.taskexecutor.TaskExecutor > - Received heartbeat request from > e88ea2f790430c9c160e540ef0546d60. > > 2020-06-24 21:10:09,286 DEBUG > org.apache.flink.runtime.taskexecutor.TaskExecutor > - Received heartbeat request from > b93d7167db364dfdcbda886944f1482f. > > 2020-06-24 21:10:09,686 INFO > org.apache.flink.runtime.taskexecutor.TaskManagerRunner > - Memory usage stats: [HEAP: 557/2573/2573 MB, NON HEAP: 111/114/424 > MB (used/committed/max)] > > 2020-06-24 21:10:09,686 INFO > org.apache.flink.runtime.taskexecutor.TaskManagerRunner > - Direct memory stats: Count: 17403, Total Capacity: 583911423, Used > Memory: 583911424 > > 2020-06-24 21:10:09,686 INFO > org.apache.flink.runtime.taskexecutor.TaskManagerRunner > - Off-heap pool stats: [Code Cache: 35/35/240 MB > (used/committed/max)], [Metaspace: 67/69/96 MB (used/committed/max)], > [Compressed Class Space: 8/9/88 MB (used/committed/max)] > > 2020-06-24 21:10:09,686 INFO > org.apache.flink.runtime.taskexecutor.TaskManagerRunner > - Garbage collector stats: [PS Scavenge, GC TIME (ms): 108643, GC > COUNT: 6981], [PS MarkSweep, GC TIME (ms): 1074, GC COUNT: 6] > > > > > 省略。。。。 > > > > > 2020-06-24 21:55:39,875 INFO > org.apache.flink.runtime.taskexecutor.TaskManagerRunner > - Direct memory stats: Count: 17403, Total Capacity: 583911423, Used > Memory: 583911424 > > 2020-06-24 21:55:39,875 INFO > org.apache.flink.runtime.taskexecutor.TaskManagerRunner > - Off-heap pool stats: [Code Cache: 35/35/240 MB > (used/committed/max)], [Metaspace: 67/69/96 MB (used/committed/max)], > [Compressed Class Space: 8/9/88 MB (used/committed/max)] > > 2020-06-24 21:55:39,876 INFO > org.apache.flink.runtime.taskexecutor.TaskManagerRunner > - Garbage collector stats: [PS Scavenge, GC TIME (ms): 110520, GC > COUNT: 7083], [PS MarkSweep, GC TIME (ms): 1074, GC COUNT: 6] > > 2020-06-24 21:55:41,721 DEBUG > org.apache.flink.streaming.runtime.tasks.StreamTask > - Starting checkpoint (316) CHECKPOINT on task Source: Custom > Source -> Map -> Filter -> Timestamps/Watermarks (4/10) > > 2020-06-24 21:55:41,721 DEBUG > org.apache.flink.runtime.state.AbstractSnapshotStrategy > - DefaultOperatorStateBackend snapshot (FsCheckpointStorageLocation > {fileSystem=org.apache.flink.core.fs.SafetyNetWrapperFileSystem@4a4b11cd, > checkpointDirectory=hdfs://xxxxchk-316, > sharedStateDirectory=hdfs://xxxxshared, > taskOwnedStateDirectory=hdfs://xxxxtaskowned, > metadataFilePath=hdfs://xxxxchk-316/_metadata, reference=(default), > fileStateSizeThreshold=1024, writeBufferSize=4096}, synchronous part) in > thread Thread[Source: Custom Source -> Map -> Filter -> > Timestamps/Watermarks (4/10),5,Flink Task Threads] took 0 ms. > > 2020-06-24 21:55:41,721 DEBUG > org.apache.flink.runtime.state.AbstractSnapshotStrategy > - DefaultOperatorStateBackend snapshot (FsCheckpointStorageLocation > {fileSystem=org.apache.flink.core.fs.SafetyNetWrapperFileSystem@4a4b11cd, > checkpointDirectory=hdfs://xxxxchk-316, > sharedStateDirectory=hdfs://xxxxshared, > taskOwnedStateDirectory=hdfs://xxxxtaskowned, > metadataFilePath=hdfs://xxxxchk-316/_metadata, reference=(default), > fileStateSizeThreshold=1024, writeBufferSize=4096}, synchronous part) in > thread Thread[Source: Custom Source -> Map -> Filter -> > Timestamps/Watermarks (4/10),5,Flink Task Threads] took 0 ms. > > 2020-06-24 21:55:41,721 DEBUG > org.apache.flink.runtime.state.AbstractSnapshotStrategy > - DefaultOperatorStateBackend snapshot (FsCheckpointStorageLocation > {fileSystem=org.apache.flink.core.fs.SafetyNetWrapperFileSystem@4a4b11cd, > checkpointDirectory=hdfs://xxxxchk-316, > sharedStateDirectory=hdfs://xxxxshared, > taskOwnedStateDirectory=hdfs://xxxxtaskowned, > metadataFilePath=hdfs://xxxxchk-316/_metadata, reference=(default), > fileStateSizeThreshold=1024, writeBufferSize=4096}, synchronous part) in > thread Thread[Source: Custom Source -> Map -> Filter -> > Timestamps/Watermarks (4/10),5,Flink Task Threads] took 0 ms. > > 2020-06-24 21:55:41,721 DEBUG > org.apache.flink.runtime.state.AbstractSnapshotStrategy > - DefaultOperatorStateBackend snapshot (FsCheckpointStorageLocation > {fileSystem=org.apache.flink.core.fs.SafetyNetWrapperFileSystem@4a4b11cd, > checkpointDirectory=hdfs://xxxxchk-316, > sharedStateDirectory=hdfs://xxxxshared, > taskOwnedStateDirectory=hdfs://xxxxtaskowned, > metadataFilePath=hdfs://xxxxchk-316/_metadata, reference=(default), > fileStateSizeThreshold=1024, writeBufferSize=4096}, synchronous part) in > thread Thread[Source: Custom Source -> Map -> Filter -> > Timestamps/Watermarks (4/10),5,Flink Task Threads] took 0 ms. > > 2020-06-24 21:55:41,721 DEBUG > org.apache.flink.streaming.runtime.tasks.StreamTask > - Finished synchronous checkpoints for checkpoint 316 on task > Source: Custom Source -> Map -> Filter -> Timestamps/Watermarks > (4/10) > > 2020-06-24 21:55:41,721 DEBUG > org.apache.flink.streaming.runtime.tasks.StreamTask > - Source: Custom Source -> Map -> Filter -> > Timestamps/Watermarks (4/10) - finished synchronous part of checkpoint 316. > Alignment duration: 0 ms, snapshot duration 0 ms > > 2020-06-24 21:55:41,737 DEBUG > org.apache.flink.runtime.state.AbstractSnapshotStrategy > - DefaultOperatorStateBackend snapshot (FsCheckpointStorageLocation > {fileSystem=org.apache.flink.core.fs.SafetyNetWrapperFileSystem@4a4b11cd, > checkpointDirectory=hdfs://xxxxchk-316, > sharedStateDirectory=hdfs://xxxxshared, > taskOwnedStateDirectory=hdfs://xxxxtaskowned, > metadataFilePath=hdfs://xxxxchk-316/_metadata, reference=(default), > fileStateSizeThreshold=1024, writeBufferSize=4096}, asynchronous part) in > thread Thread[AsyncOperations-thread-316,5,Flink Task Threads] took 16 ms. > > 2020-06-24 21:55:41,738 DEBUG > org.apache.flink.streaming.runtime.tasks.StreamTask > - Source: Custom Source -> Map -> Filter -> > Timestamps/Watermarks (4/10) - finished asynchronous part of checkpoint > 316. Asynchronous duration: 16 ms > > 2020-06-24 21:55:42,008 DEBUG > org.apache.flink.streaming.runtime.io.CheckpointBarrierAligner > - Window(SlidingEventTimeWindows(10800000, 300000, 0), EventTimeTrigger, > anon$11, ScalaProcessWindowFunctionWrapper) -> Map (10/10) > (a608e8a1ab75622dc57f164bdbb86743): Received barrier from channel 3. > > 2020-06-24 21:55:42,008 DEBUG > org.apache.flink.streaming.runtime.io.CheckpointBarrierAligner > - Window(SlidingEventTimeWindows(10800000, 300000, 0), EventTimeTrigger, > anon$11, ScalaProcessWindowFunctionWrapper) -> Map (10/10) > (a608e8a1ab75622dc57f164bdbb86743): Starting stream alignment for > checkpoint 316. > > 2020-06-24 21:55:42,110 DEBUG > org.apache.flink.streaming.runtime.io.CheckpointBarrierAligner > - Window(SlidingEventTimeWindows(10800000, 300000, 0), EventTimeTrigger, > anon$11, ScalaProcessWindowFunctionWrapper) -> Map (10/10) > (a608e8a1ab75622dc57f164bdbb86743): Received barrier from channel 7. > > 2020-06-24 21:55:42,110 DEBUG > org.apache.flink.streaming.runtime.io.CheckpointBarrierAligner > - Window(SlidingEventTimeWindows(10800000, 300000, 0), EventTimeTrigger, > anon$11, ScalaProcessWindowFunctionWrapper) -> Map (10/10) > (a608e8a1ab75622dc57f164bdbb86743): Received barrier from channel 2. > > 2020-06-24 21:55:42,110 DEBUG > org.apache.flink.streaming.runtime.io.CheckpointBarrierAligner > - Window(SlidingEventTimeWindows(10800000, 300000, 0), EventTimeTrigger, > anon$11, ScalaProcessWindowFunctionWrapper) -> Map (10/10) > (a608e8a1ab75622dc57f164bdbb86743): Received barrier from channel 8. > > 2020-06-24 21:55:42,110 DEBUG > org.apache.flink.streaming.runtime.io.CheckpointBarrierAligner > - Window(SlidingEventTimeWindows(10800000, 300000, 0), EventTimeTrigger, > anon$11, ScalaProcessWindowFunctionWrapper) -> Map (10/10) > (a608e8a1ab75622dc57f164bdbb86743): Received barrier from channel 6. > > 2020-06-24 21:55:42,110 DEBUG > org.apache.flink.streaming.runtime.io.CheckpointBarrierAligner > - Window(SlidingEventTimeWindows(10800000, 300000, 0), EventTimeTrigger, > anon$11, ScalaProcessWindowFunctionWrapper) -> Map (10/10) > (a608e8a1ab75622dc57f164bdbb86743): Received barrier from channel 0. > > 2020-06-24 21:55:42,110 DEBUG > org.apache.flink.streaming.runtime.io.CheckpointBarrierAligner > - Window(SlidingEventTimeWindows(10800000, 300000, 0), EventTimeTrigger, > anon$11, ScalaProcessWindowFunctionWrapper) -> Map (10/10) > (a608e8a1ab75622dc57f164bdbb86743): Received barrier from channel 5. > > 2020-06-24 21:55:42,110 DEBUG > org.apache.flink.streaming.runtime.io.CheckpointBarrierAligner > - Window(SlidingEventTimeWindows(10800000, 300000, 0), EventTimeTrigger, > anon$11, ScalaProcessWindowFunctionWrapper) -> Map (10/10) > (a608e8a1ab75622dc57f164bdbb86743): Received barrier from channel 9. > > 2020-06-24 21:55:42,110 DEBUG > org.apache.flink.streaming.runtime.io.CheckpointBarrierAligner > - Window(SlidingEventTimeWindows(10800000, 300000, 0), EventTimeTrigger, > anon$11, ScalaProcessWindowFunctionWrapper) -> Map (10/10) > (a608e8a1ab75622dc57f164bdbb86743): Received barrier from channel 4. > > 2020-06-24 21:55:42,110 DEBUG > org.apache.flink.streaming.runtime.io.CheckpointBarrierAligner > - Window(SlidingEventTimeWindows(10800000, 300000, 0), EventTimeTrigger, > anon$11, ScalaProcessWindowFunctionWrapper) -> Map (10/10) > (a608e8a1ab75622dc57f164bdbb86743): Received barrier from channel 1. > > 2020-06-24 21:55:42,110 DEBUG > org.apache.flink.streaming.runtime.io.CheckpointBarrierAligner > - Window(SlidingEventTimeWindows(10800000, 300000, 0), EventTimeTrigger, > anon$11, ScalaProcessWindowFunctionWrapper) -> Map (10/10) > (a608e8a1ab75622dc57f164bdbb86743): Received all barriers, triggering > checkpoint 316 at 1593004193363. > > 2020-06-24 21:55:42,110 DEBUG > org.apache.flink.streaming.runtime.io.CheckpointBarrierAligner > - Window(SlidingEventTimeWindows(10800000, 300000, 0), EventTimeTrigger, > anon$11, ScalaProcessWindowFunctionWrapper) -> Map (10/10) > (a608e8a1ab75622dc57f164bdbb86743): End of stream alignment, feeding > buffered data back. > > 2020-06-24 21:55:42,110 DEBUG > org.apache.flink.streaming.runtime.tasks.StreamTask > - Starting checkpoint (316) CHECKPOINT on task > Window(SlidingEventTimeWindows(10800000, 300000, 0), EventTimeTrigger, > anon$11, ScalaProcessWindowFunctionWrapper) -> Map (10/10) > > 2020-06-24 21:55:42,110 DEBUG > org.apache.flink.runtime.state.AbstractSnapshotStrategy > - DefaultOperatorStateBackend snapshot (FsCheckpointStorageLocation > {fileSystem=org.apache.flink.core.fs.SafetyNetWrapperFileSystem@61164119, > checkpointDirectory=hdfs://xxxxchk-316, > sharedStateDirectory=hdfs://xxxxshared, > taskOwnedStateDirectory=hdfs://xxxxtaskowned, > metadataFilePath=hdfs://xxxxchk-316/_metadata, reference=(default), > fileStateSizeThreshold=1024, writeBufferSize=4096}, synchronous part) in > thread Thread[Window(SlidingEventTimeWindows(10800000, 300000, 0), > EventTimeTrigger, anon$11, ScalaProcessWindowFunctionWrapper) -> Map > (10/10),5,Flink Task Threads] took 0 ms. > > 2020-06-24 21:55:42,110 DEBUG > org.apache.flink.runtime.state.AbstractSnapshotStrategy > - DefaultOperatorStateBackend snapshot (FsCheckpointStorageLocation > {fileSystem=org.apache.flink.core.fs.SafetyNetWrapperFileSystem@61164119, > checkpointDirectory=hdfs://xxxxchk-316, > sharedStateDirectory=hdfs://xxxxshared, > taskOwnedStateDirectory=hdfs://xxxxtaskowned, > metadataFilePath=hdfs://xxxxchk-316/_metadata, reference=(default), > fileStateSizeThreshold=1024, writeBufferSize=4096}, synchronous part) in > thread Thread[Window(SlidingEventTimeWindows(10800000, 300000, 0), > EventTimeTrigger, anon$11, ScalaProcessWindowFunctionWrapper) -> Map > (10/10),5,Flink Task Threads] took 0 ms. > > 2020-06-24 21:55:42,980 DEBUG > org.apache.flink.runtime.state.AbstractSnapshotStrategy > - Asynchronous incremental RocksDB snapshot > (FsCheckpointStorageLocation > {fileSystem=org.apache.flink.core.fs.SafetyNetWrapperFileSystem@61164119, > checkpointDirectory=hdfs://xxxx/chk-316, > sharedStateDirectory=hdfs://xxxx/shared, > taskOwnedStateDirectory=hdfs://xxxx/taskowned, > metadataFilePath=hdfs://xxxx/chk-316/_metadata, reference=(default), > fileStateSizeThreshold=1024, writeBufferSize=4096}, synchronous part) in > thread Thread[Window(SlidingEventTimeWindows(10800000, 300000, 0), > EventTimeTrigger, anon$11, ScalaProcessWindowFunctionWrapper) -> Map > (10/10),5,Flink Task Threads] took 870 ms. > > 2020-06-24 21:55:42,981 DEBUG > org.apache.flink.streaming.runtime.tasks.StreamTask > - Finished synchronous checkpoints for checkpoint 316 on task > Window(SlidingEventTimeWindows(10800000, 300000, 0), EventTimeTrigger, > anon$11, ScalaProcessWindowFunctionWrapper) -> Map (10/10) > > 2020-06-24 21:55:42,981 DEBUG > org.apache.flink.streaming.runtime.tasks.StreamTask > - Window(SlidingEventTimeWindows(10800000, 300000, 0), > EventTimeTrigger, anon$11, ScalaProcessWindowFunctionWrapper) -> Map > (10/10) - finished synchronous part of checkpoint 316. Alignment duration: > 101 ms, snapshot duration 870 ms > > 2020-06-24 21:55:42,981 DEBUG > org.apache.flink.streaming.runtime.io.CachedBufferStorage - > Window(SlidingEventTimeWindows(10800000, 300000, 0), EventTimeTrigger, > anon$11, ScalaProcessWindowFunctionWrapper) -> Map (10/10) > (a608e8a1ab75622dc57f164bdbb86743): Size of buffered data: 98304 bytes > > 2020-06-24 21:55:42,981 DEBUG > org.apache.flink.streaming.runtime.io.CachedBufferStorage - > Window(SlidingEventTimeWindows(10800000, 300000, 0), EventTimeTrigger, > anon$11, ScalaProcessWindowFunctionWrapper) -> Map (10/10) > (a608e8a1ab75622dc57f164bdbb86743): Finished feeding back buffered data. > > 2020-06-24 21:55:43,814 DEBUG > org.apache.flink.runtime.taskexecutor.TaskExecutor > - Received heartbeat request from > e88ea2f790430c9c160e540ef0546d60. > > 2020-06-24 21:55:44,758 DEBUG > org.apache.flink.runtime.taskexecutor.TaskExecutor > - Received heartbeat request from > b93d7167db364dfdcbda886944f1482f. > > 2020-06-24 21:55:45,714 DEBUG > org.apache.flink.runtime.state.AbstractSnapshotStrategy > - Asynchronous incremental RocksDB snapshot > (FsCheckpointStorageLocation > {fileSystem=org.apache.flink.core.fs.SafetyNetWrapperFileSystem@61164119, > checkpointDirectory=hdfs://xxxx/chk-316, > sharedStateDirectory=hdfs://xxxx/shared, > taskOwnedStateDirectory=hdfs://xxxx/taskowned, > metadataFilePath=hdfs://xxxx/chk-316/_metadata, reference=(default), > fileStateSizeThreshold=1024, writeBufferSize=4096}, asynchronous part) in > thread Thread[AsyncOperations-thread-316,5,Flink Task Threads] took 2733 ms. > > 2020-06-24 21:55:45,715 DEBUG > org.apache.flink.streaming.runtime.tasks.StreamTask > - Window(SlidingEventTimeWindows(10800000, 300000, 0), > EventTimeTrigger, anon$11, ScalaProcessWindowFunctionWrapper) -> Map > (10/10) - finished asynchronous part of checkpoint 316. Asynchronous > duration: 2734 ms > > 2020-06-24 21:55:49,876 INFO > org.apache.flink.runtime.taskexecutor.TaskManagerRunner > - Memory usage stats: [HEAP: 220/2589/2589 MB, NON HEAP: 111/114/424 > MB (used/committed/max)] > > 2020-06-24 21:55:49,876 INFO > org.apache.flink.runtime.taskexecutor.TaskManagerRunner > - Direct memory stats: Count: 17414, Total Capacity: 584128352, Used > Memory: 584128353 > > 2020-06-24 21:55:49,876 INFO > org.apache.flink.runtime.taskexecutor.TaskManagerRunner > - Off-heap pool stats: [Code Cache: 35/35/240 MB > (used/committed/max)], [Metaspace: 67/69/96 MB (used/committed/max)], > [Compressed Class Space: 8/9/88 MB (used/committed/max)] > > 2020-06-24 21:55:49,876 INFO > org.apache.flink.runtime.taskexecutor.TaskManagerRunner > - Garbage collector stats: [PS Scavenge, GC TIME (ms): 110525, GC > COUNT: 7084], [PS MarkSweep, GC TIME (ms): 1074, GC COUNT: 6] > > > > > > -- > ************************************** > tivanli > ************************************** -- ************************************** tivanli **************************************
