JVM运维调优常用命令收集(运维及开发通用)！！！！

jps

查看当前机器上所有运行的java进程名称与pid(进程编号)

jps -l

[root@flink001 ~]# jps -l
24976 ossclient-0.0.1-SNAPSHOT.jar
6560 org.apache.hadoop.hdfs.server.namenode.NameNode
12800 org.apache.flink.runtime.taskexecutor.TaskManagerRunner
13425 sun.tools.jps.Jps
6724 org.apache.hadoop.hdfs.server.datanode.DataNode
26460 org.apache.zookeeper.server.quorum.QuorumPeerMain
8125 kafka.Kafka
6943 org.apache.hadoop.hdfs.server.namenode.SecondaryNameNode

jmap

导出java进程堆命令

jmap -dump:format=b,file=outfile pid

-histo[:live] 打印每个class的实例数目,内存占用,类全名信息. VM的内部类名字开头会加上前缀”*”. 如果live子参数加上后,只统计活的对象数量.

jmap -histo:live pid
结果：
[root@flink001 ~]# jmap -histo:live 8125 |head -n100

 num     #instances         #bytes  class name
----------------------------------------------
   1:           775      135346896  [B
   2:         25867        2301232  [C
   3:         54297        1303128  java.util.concurrent.ConcurrentSkipListMap$Node
   4:         53879        1293096  java.lang.Double
   5:          6920         771264  java.lang.Class
   6:         26814         643536  java.util.concurrent.ConcurrentSkipListMap$Index
   7:         25791         618984  java.lang.String
   8:         12794         409408  java.util.concurrent.ConcurrentHashMap$Node
   9:         10845         347040  java.util.HashMap$Node
  10:          4784         306176  java.util.concurrent.ConcurrentHashMap
  11:          4496         251840  [Ljava.lang.Object;
  12:          7713         185112  java.lang.Long
  13:         10860         173760  java.lang.Object
  14:          1350         168600  [Ljava.util.HashMap$Node;
  15:          4981         159392  java.util.concurrent.locks.ReentrantLock$NonfairSync
  16:          4499         143968  scala.collection.immutable.Map$Map2
  17:          5553         133272  scala.collection.mutable.DefaultEntry
  18:          4696         112704  kafka.utils.Pool
  19:          4608         110592  kafka.server.DelayedOperationPurgatory$WatcherList
  20:           164         108960  [Ljava.util.concurrent.ConcurrentHashMap$Node;
  21:          4450         106800  kafka.network.RequestMetrics$ErrorMeter
  22:          1831          96312  [I
  23:          3821          91704  java.util.concurrent.atomic.AtomicLong

jinfo

查看指定的jvm进程所有的属性设置和配置参数

jinfo pid

jmap

查看某个pid进程对应的应用程序内存占用情况

jmap -heap pid

[self@owen-first ~]# jmap -heap 3702
Attaching to process ID 3702, please wait...
Debugger attached successfully.
Server compiler detected.
JVM version is 25.161-b12

using thread-local object allocation.
Mark Sweep Compact GC

Heap Configuration:
   MinHeapFreeRatio         = 40
   MaxHeapFreeRatio         = 70
   MaxHeapSize              = 268435456 (256.0MB)
   NewSize                  = 67108864 (64.0MB)
   MaxNewSize               = 134217728 (128.0MB)
   OldSize                  = 67108864 (64.0MB)
   NewRatio                 = 1
   SurvivorRatio            = 8
   MetaspaceSize            = 21807104 (20.796875MB)
   CompressedClassSpaceSize = 1073741824 (1024.0MB)
   MaxMetaspaceSize         = 17592186044415 MB
   G1HeapRegionSize         = 0 (0.0MB)

Heap Usage:
New Generation (Eden + 1 Survivor Space):
   capacity = 70582272 (67.3125MB)
   used     = 47699728 (45.49000549316406MB)
   free     = 22882544 (21.822494506835938MB)
   67.58032385242572% used
Eden Space:
   capacity = 62783488 (59.875MB)
   used     = 43811552 (41.781951904296875MB)
   free     = 18971936 (18.093048095703125MB)
   69.78196560216597% used
From Space:
   capacity = 7798784 (7.4375MB)
   used     = 3888176 (3.7080535888671875MB)
   free     = 3910608 (3.7294464111328125MB)
   49.85618270745798% used
To Space:
   capacity = 7798784 (7.4375MB)
   used     = 0 (0.0MB)
   free     = 7798784 (7.4375MB)
   0.0% used
tenured generation:
   capacity = 78106624 (74.48828125MB)
   used     = 53877384 (51.38147735595703MB)
   free     = 24229240 (23.10680389404297MB)
   68.97927632872725% used

31332 interned Strings occupying 3263472 bytes.

jstack

查看进程所包含所有线程的Java堆栈信息

[self@owen-first ~]# jstack 3702
2019-11-14 14:23:22
Full thread dump Java HotSpot(TM) 64-Bit Server VM (25.161-b12 mixed mode):

"Attach Listener" #1743 daemon prio=9 os_prio=0 tid=0x00007f13600ac000 nid=0x5533 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"pool-3-thread-37" #1688 prio=5 os_prio=0 tid=0x00007f135ca08000 nid=0x51eb waiting on condition [0x00007f1354c8b000]
   java.lang.Thread.State: WAITING (parking)
	at sun.misc.Unsafe.park(Native Method)
	- parking to wait for  <0x00000000f4356e20> (a java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject)
	at java.util.concurrent.locks.LockSupport.park(LockSupport.java:175)
	at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await(AbstractQueuedSynchronizer.java:2039)
	at java.util.concurrent.LinkedBlockingQueue.take(LinkedBlockingQueue.java:442)
	at run.halo.app.event.post.AbstractVisitEventListener$PostVisitTask.run(AbstractVisitEventListener.java:101)
	at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
	at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
	at java.lang.Thread.run(Thread.java:748)

......

"Reference Handler" #2 daemon prio=10 os_prio=0 tid=0x00007f138809c800 nid=0xe79 in Object.wait() [0x00007f138e3bd000]
   java.lang.Thread.State: WAITING (on object monitor)
	at java.lang.Object.wait(Native Method)
	at java.lang.Object.wait(Object.java:502)
	at java.lang.ref.Reference.tryHandlePending(Reference.java:191)
	- locked <0x00000000f802b410> (a java.lang.ref.Reference$Lock)
	at java.lang.ref.Reference$ReferenceHandler.run(Reference.java:153)

"VM Thread" os_prio=0 tid=0x00007f1388095000 nid=0xe78 runnable 

"VM Periodic Task Thread" os_prio=0 tid=0x00007f13880e0800 nid=0xe7f waiting on condition 

JNI global references: 2336

jstat

可以实时监测系统资源占用与jvm运行情况，全称：“Java Virtual Machine statistics monitoring tool”。

// 命令语法结构：
Usage: jstat -help|-options
  jstat -<option> [-t] [-h<lines>] <vmid> [<interval> [<count>]]
 
// 参数解释：
Options — 选项，我们一般使用 -gcutil 查看gc情况
vmid — VM的进程号，即当前运行的java进程号
interval– 间隔时间，单位为秒或者毫秒
count — 打印次数，如果缺省则打印无数次
 
S0 — Heap上的 Survivor space 0 区已使用空间的百分比
S1 — Heap上的 Survivor space 1 区已使用空间的百分比
E — Heap上的 Eden space 区已使用空间的百分比
O — Heap上的 Old space 区已使用空间的百分比
P — Perm space 区已使用空间的百分比
YGC — 从应用程序启动到采样时发生 Young GC 的次数
YGCT– 从应用程序启动到采样时 Young GC 所用的时间(单位秒)
FGC — 从应用程序启动到采样时发生 Full GC 的次数
FGCT– 从应用程序启动到采样时 Full GC 所用的时间(单位秒)
GCT — 从应用程序启动到采样时用于垃圾回收的总时间(单位秒)
 
//示例
jstat -options
-class
-compiler
-gc
-gccapacity
-gccause
-gcmetacapacity
-gcnew
-gcnewcapacity
-gcold
-gcoldcapacity
-gcutil
-printcompilation
 
[self@owen-first ~]# jstat -class -t 3702
Timestamp       Loaded  Bytes  Unloaded  Bytes     Time   
       176180.9  17036 30914.4      211   302.6      21.65

jconsole

以GUI的方式更直观化呈现jvm进程的实时情况，比如内存占用, 线程执行情况等

在jdk_home/bin目录下执行 jconsole.exe 打开图形化界面

jprofiler

一个付费的商业jvm分析监控工具, 可查看概况, 内存, GC活动, class 状况, 线程信息, CPU 占用情况, 内存对象信息, 还有数据库连接等, 是一个非常优秀的分析工具