[100 cases of JVM tuning practice] 03 -- four cases of JVM heap tuning

front said
Author's brief introduction ： Half old 518, Long distance runner , Determined to persist in writing 10 Blog of the year , Focus on java Back end
Column Introduction ： Case driven introduction JVM knowledge , Teach you how to use JVM Troubleshooting 、 Evaluation code 、 Optimize performance
The article brief introduction ： Introduce the related concepts of heap 、 Teach you to check 5 A common JVM Pile of cases

6. Pile up

6.1 Characteristics of reactor

Use new All objects created by keyword will use heap .

characteristic ：

• Thread sharing , The objects in the heap need to consider thread safety .
• There's a garbage collection mechanism .

6.2 Heap memory overflow problem

There is a garbage collection mechanism in the heap , But the premise of garbage collection is that the objects in the heap are no longer referenced ( actually , The algorithm for recycling references is the root reachable Algorithm , I'll talk about it later , The statement here is inaccurate ), So if we have too many objects that cannot be recycled , It may cause memory overflow .

public class MemoryOverFlow {
    
    public static void main(String[] args) {
    
        int i = 0;
        String a = "hello";
        List list = new ArrayList(); //  until catch Code block execution , Has been used 
        try {
    
            while (true) {
    
                list.add(a);
                a = a + a;
                i++;
            }
        } catch (Throwable e) {
     // Use Throwable, If you use Exception I can't wrap it up Error,i Can't be printed out 
           e.printStackTrace();
           System.out.println(i);
        }

    }
}

appear OutOfMemoryError

java.lang.OutOfMemoryError: Overflow: String length out of range
        at java.base/java.lang.StringConcatHelper.checkOverflow(StringConcatHelper.java:57)
        at java.base/java.lang.StringConcatHelper.mix(StringConcatHelper.java:138)
        at java.base/java.lang.StringConcatHelper.simpleConcat(StringConcatHelper.java:420)
        at MemoryOverFlow.main(MemoryOverFlow.java:12)
28

another ： Parameters -Xmx You can set jvm Memory size , When troubleshooting heap memory problems, you can set it to be small ( Such as 8m), More likely to expose memory overflow problems . Setup method ： Click on build and run Go with you modify options->add vm options.

6.3 Evaluation of the impact of code memory performance

I wrote a piece of code in my work , How to judge the impact of a piece of code on memory performance ？ You can use the following tools .

• jps See what the system has java process

• jmap Check the occupation of heap memory at a certain time

• jconsole Multifunctional real-time monitoring tool

Through the following demo To demonstrate .

public class jvmdemo {
    
	public static void main(String[] args) throws InterruptedException {
    
        System.out.println("1....."); // Output hint , Easy to carry out Heap Dump
        Thread.sleep(60000); // to 30s Time is used for Heap Dump
        byte [] arr = new byte[1024 * 1024 * 10];
        System.out.println("2.......");
        Thread.sleep(60000);
        arr = null;
        System.gc();
        System.out.println("3......");
        Thread.sleep(100000L);
	}

}

When the output 1… after , Execute first jps see jvmdemo Corresponding pid

tip： If you are windows System ,jps No results returned , You can refer to the blog

Windows in jps The command cannot be viewed java Process issues _ countless _mirage The blog of -CSDN Blog _windows see java process

give the result as follows .

perform jmap -heap xxx(pid) Check the heap memory usage at this time .

tip:

If you execute jmap Report errors .

Error: -heap option used
Cannot connect to core dump or remote debug server. Use jhsdb jmap instead

Because jdk8 Later versions, previous jmap -heap xxx(pid) The command can no longer be used . You can use the command instead jhsdb jmap --heap --pid xxx.

In the prompt message output 1,2,3 After three operations, the results are as follows .

Heap Usage:
G1 Heap:
   regions  = 2034
   capacity = 8531214336 (8136.0MB)
   used     = 0 (0.0MB)
   free     = 8531214336 (8136.0MB)
   0.0% used
G1 Young Generation:
Eden Space:
   regions  = 0
   capacity = 29360128 (28.0MB)
   used     = 0 (0.0MB)
   free     = 29360128 (28.0MB)
   0.0% used
Survivor Space:
   regions  = 0
   capacity = 0 (0.0MB)
   used     = 0 (0.0MB)
   free     = 0 (0.0MB)
   0.0% used
G1 Old Generation:
   regions  = 0
   capacity = 507510784 (484.0MB)
   used     = 0 (0.0MB)
   free     = 507510784 (484.0MB)
   0.0% used

Heap Usage:
G1 Heap:
   regions  = 2034
   capacity = 8531214336 (8136.0MB)
   used     = 12582912 (12.0MB)
   free     = 8518631424 (8124.0MB)
   0.14749262536873156% used
G1 Young Generation:
Eden Space:
   regions  = 0
   capacity = 29360128 (28.0MB)
   used     = 0 (0.0MB)
   free     = 29360128 (28.0MB)
   0.0% used
Survivor Space:
   regions  = 0
   capacity = 0 (0.0MB)
   used     = 0 (0.0MB)
   free     = 0 (0.0MB)
   0.0% used
G1 Old Generation:
   regions  = 3
   capacity = 507510784 (484.0MB)
   used     = 12582912 (12.0MB)
   free     = 494927872 (472.0MB)
   2.479338842975207% used

Heap Usage:
G1 Heap:
   regions  = 2034
   capacity = 8531214336 (8136.0MB)
   used     = 673872 (0.6426544189453125MB)
   free     = 8530540464 (8135.357345581055MB)
   0.007898898954588403% used
G1 Young Generation:
Eden Space:
   regions  = 0
   capacity = 8388608 (8.0MB)
   used     = 0 (0.0MB)
   free     = 8388608 (8.0MB)
   0.0% used
Survivor Space:
   regions  = 0
   capacity = 0 (0.0MB)
   used     = 0 (0.0MB)
   free     = 0 (0.0MB)
   0.0% used
G1 Old Generation:
   regions  = 1
   capacity = 8388608 (8.0MB)
   used     = 673872 (0.6426544189453125MB)
   free     = 7714736 (7.3573455810546875MB)
   8.033180236816406% used

Focus on uesd This one , You can see the changing process of memory in the code , here JVM version yes 16.0.2+7-67, Different versions may vary slightly .

Use jconsole Data can be observed in real time , And it's not just about observing local processes , You can also observe remote processes .

The method is simple to use . Enter... At the command terminal jconsole It will pop up automatically .

[ Failed to transfer the external chain picture , The origin station may have anti-theft chain mechanism , It is suggested to save the pictures and upload them directly (img-27AjN2Zs-1656678910764)(F:/%E5%8D%9A%E5%AE%A2%E5%9B%BE%E7%89%87/jvm/image-20220630213435635.png)]

The observations of the above code are as follows .

summary ：

The above is for troubleshooting , We add log as well as sleep The way , Help us with dump Grasp of time , Similar methods can be used in actual production .

Except for memory ,jconsole You can also monitor threads 、cpu Occupancy rate and the number of classes change .

[ Failed to transfer the external chain picture , The origin station may have anti-theft chain mechanism , It is suggested to save the pictures and upload them directly (img-J3O31REj-1656678910765)(F:/%E5%8D%9A%E5%AE%A2%E5%9B%BE%E7%89%87/jvm/image-20220630214317498.png)]

It can also help us detect deadlocks .

[ Failed to transfer the external chain picture , The origin station may have anti-theft chain mechanism , It is suggested to save the pictures and upload them directly (img-98hyGDJs-1656678910765)(F:/%E5%8D%9A%E5%AE%A2%E5%9B%BE%E7%89%87/jvm/image-20220630214429419.png)]

6.4 The memory occupation of multiple garbage collection is still very high, and the troubleshooting of the problem

jvisualvm It's also a visualization tool , Than jconsole Better to use .

The method is simple to use , Use command which java Find the java The installation path , And switch to this path , Then execute on the corresponding path jvisualvm that will do .

In the high version JDK（ Greater than 1.8 Or later updated 1.8 edition ） It will no longer be automatically integrated in . The reference blog can download the independent version ：JDK There is no higher version VisualVM_ Dongli Jike's blog -CSDN Blog .

Now use it to demonstrate , The memory occupation of multiple garbage collection is still very high, and the troubleshooting of the problem

/** *  Show me how many objects to view   Heap dump  dump */
public class Demo1_13 {
    

    public static void main(String[] args) throws InterruptedException {
    
        List<Student> students = new ArrayList<>();
        for (int i = 0; i < 200; i++) {
    
            students.add(new Student());
// Student student = new Student();
        }
        Thread.sleep(1000000000L);
    }
}
class Student {
    
    private byte[] big = new byte[1024*1024];
}

Use jvisualvm Click to select the corresponding process , You can see that the heap memory is very high , Carry out garbage collection .

[ Failed to transfer the external chain picture , The origin station may have anti-theft chain mechanism , It is suggested to save the pictures and upload them directly (img-gCeFjrEJ-1656678910765)(F:/%E5%8D%9A%E5%AE%A2%E5%9B%BE%E7%89%87/jvm/image-20220630215736401.png)]

Memory has not been reduced much .

Conduct Heap Dump, Select the right menu bar , According to the occupied memory class Sort

[ Failed to transfer the external chain picture , The origin station may have anti-theft chain mechanism , It is suggested to save the pictures and upload them directly (img-OXUm6CVo-1656678910766)(F:/%E5%8D%9A%E5%AE%A2%E5%9B%BE%E7%89%87/jvm/image-20220630215949114.png)]

According to the occupied memory class The sorting results are as follows .

Click the class to see all instances of the corresponding class , We click on the one that occupies the most memory ArrayList Check out the examples , Click elementData You can see the details .

The discovery may be student The pot , You can see some examples , Confirm the conclusion .

Point again student object , Navigate to the property big. It is found that an attribute takes about 1M Space . I can't stand it

Then locate the corresponding source code .

/** *  Show me how many objects to view   Heap dump  dump */
public class Demo1_13 {
    

    public static void main(String[] args) throws InterruptedException {
    
        List<Student> students = new ArrayList<>();
        for (int i = 0; i < 200; i++) {
    
            students.add(new Student());
// Student student = new Student();
        }
        Thread.sleep(1000000000L);
    }
}
class Student {
    
    private byte[] big = new byte[1024*1024];
}

Sure enough , We have 200 individual Student object , Holds the memory footprint 1M Of big example 200 individual , And has been living .