Flink总结-设置Jvm参数

迈不过友情╰ 2023-07-13 13:48 20阅读 0赞

## jvm设置 ##

[http://www.360doc.com/content/12/1023/16/9615799\_243296263.shtml][http_www.360doc.com_content_12_1023_16_9615799_243296263.shtml]

堆设置
    -Xms :初始堆大小
    -Xmx :最大堆大小
    -XX:NewSize=n :设置年轻代大小
    -XX:NewRatio=n: 设置年轻代和年老代的比值。如:为3，表示年轻代与年老代比值为1：3，年轻代占整个年轻代年老代和的1/4
    -XX:SurvivorRatio=n :年轻代中Eden区与两个Survivor区的比值。注意Survivor区有两个。如：3，表示Eden：Survivor=3：2，一个Survivor区占整个年轻代的1/5
    -XX:MaxPermSize=n :设置持久代大小
    收集器设置
    -XX:+UseSerialGC :设置串行收集器
    -XX:+UseParallelGC :设置并行收集器
    -XX:+UseParalledlOldGC :设置并行年老代收集器
    -XX:+UseConcMarkSweepGC :设置并发收集器
    垃圾回收统计信息
    -XX:+PrintHeapAtGC GC的heap详情
    -XX:+PrintGCDetails  GC详情
    -XX:+PrintGCTimeStamps  打印GC时间信息
    -XX:+PrintTenuringDistribution    打印年龄信息等
    -XX:+HandlePromotionFailure   老年代分配担保（true  or false）
    并行收集器设置
    -XX:ParallelGCThreads=n :设置并行收集器收集时使用的CPU数。并行收集线程数。
    -XX:MaxGCPauseMillis=n :设置并行收集最大暂停时间
    -XX:GCTimeRatio=n :设置垃圾回收时间占程序运行时间的百分比。公式为1/(1+n)
    并发收集器设置
    -XX:+CMSIncrementalMode :设置为增量模式。适用于单CPU情况。
    -XX:ParallelGCThreads=n :设置并发收集器年轻代收集方式为并行收集时，使用的CPU数。并行收集线程数

## flink设置jvm参数 ##

[https://ci.apache.org/projects/flink/flink-docs-release-1.3/setup/config.html][https_ci.apache.org_projects_flink_flink-docs-release-1.3_setup_config.html]

./flink-1.2.0/bin/flink run -m yarn-cluster -yn 4 -yjm 2048 -ytm 8086 -c beam.count.OpsCount -yqu data-default \
    -yD taskmanager.heap.mb=4096 -yD yarn.heap-cutoff-ratio=0.6 -yD taskmanager.debug.memory.startLogThread=true -yD taskmanager.debug.memory.logIntervalMs=600000 \
    -yz toratest -yst -yd ./beampoc-bundled-0.0.1-SNAPSHOT.jar --parallelism=4

[https://mingminxu.com/2017/01/16/run-flink-jobs-on-yarn/][https_mingminxu.com_2017_01_16_run-flink-jobs-on-yarn]  
  
增加flink的jvm参数

-yD env.java.opts="-XX:NewRatio=2"

## 默认flink 4container slot4 的jmap输出 ##

Heap Configuration:
       MinHeapFreeRatio         = 0
       MaxHeapFreeRatio         = 100
       MaxHeapSize              = 2359296000 (2250.0MB)
       NewSize                  = 786432000 (750.0MB)
       MaxNewSize               = 786432000 (750.0MB)
       OldSize                  = 1572864000 (1500.0MB)
       NewRatio                 = 2
       SurvivorRatio            = 8
       MetaspaceSize            = 21807104 (20.796875MB)
       CompressedClassSpaceSize = 1073741824 (1024.0MB)
       MaxMetaspaceSize         = 17592186044415 MB
       G1HeapRegionSize         = 0 (0.0MB)
    
    Heap Usage:
    PS Young Generation
    Eden Space:
       capacity = 476053504 (454.0MB)
       used     = 224818576 (214.40370178222656MB)
       free     = 251234928 (239.59629821777344MB)
       47.225484974058716% used
    From Space:
       capacity = 101187584 (96.5MB)
       used     = 85147536 (81.20301818847656MB)
       free     = 16040048 (15.296981811523438MB)
       84.14820537665965% used
    To Space:
       capacity = 155189248 (148.0MB)
       used     = 0 (0.0MB)
       free     = 155189248 (148.0MB)
       0.0% used
    PS Old Generation
       capacity = 1572864000 (1500.0MB)
       used     = 922110224 (879.3928375244141MB)
       free     = 650753776 (620.6071624755859MB)
       58.62618916829427% used
    
    22206 interned Strings occupying 2255248 bytes.

0人点赞

Flink-流计算

作者：zachary\_1db5  
链接：https://www.jianshu.com/p/6f2f53358b9c  
来源：简书  
著作权归作者所有。商业转载请联系作者获得授权，非商业转载请注明出处。

[http_www.360doc.com_content_12_1023_16_9615799_243296263.shtml]: https://link.jianshu.com?t=http%3A%2F%2Fwww.360doc.com%2Fcontent%2F12%2F1023%2F16%2F9615799_243296263.shtml
[https_ci.apache.org_projects_flink_flink-docs-release-1.3_setup_config.html]: https://link.jianshu.com?t=https%3A%2F%2Fci.apache.org%2Fprojects%2Fflink%2Fflink-docs-release-1.3%2Fsetup%2Fconfig.html
[https_mingminxu.com_2017_01_16_run-flink-jobs-on-yarn]: https://link.jianshu.com?t=https%3A%2F%2Fmingminxu.com%2F2017%2F01%2F16%2Frun-flink-jobs-on-yarn%2F