Author Archives: Robins

[Solved] Chrome Error: The request client is not a secure context

1. Background

When using Ajax to access across domains, an error is suddenly reported:

 Access to XMLHttpRequest at '' from origin 'http://xxx:yyy' 
 has been blocked by CORS policy: 
 The request client is not a secure context 
 and the resource is in more-private address space `local`.

The general meaning is that the request is blocked by cross-domain policy, and the client is not secure.

2. Analysis

Baidu has found many times that it is caused by the upgrade of Chrome browser, and this problem also exists in edge browser.

3. Solution

Enter in the address field chrome://flags/#block -Secure private network requests , change the option to disable disable, and then restart the browser

[Solved] Springboot upload failed to find the temporary directory error

The springboot upload failed to find the temporary directory and reported an error

1. Problem description

According to the feedback of online users, the upload file function suddenly reported an error. After checking the log file, the error message is:
failed to parse multipart servlet request; nested exception is java.lang.RuntimeException: java.nio.file.NoSuchFileException: /tmp/undertow.8099.1610110889708131476/undertow1171565914461190366upload

2. Cause investigation

Originally, when launching the spring boot application through java - jar in the linux operating system, a temporary directory will be created by default in the /tmp directory (in Windows operating system, C:\users\default\appdata\ local\temp), and the temporary directory is generally in the format of undertow. Port.* (if the Tomcat container is Tomcat.Port.*, this article will take undertow as an example, and Tomcat is the same.) , files need to be converted into temporary files and stored here when uploading. However, if the files in the /tmp directory are not used for more than 10 days, they will be automatically cleaned up by the system. Therefore, the above problems do not occur in the directory when uploading again.

3. Problem recurrence

Since the temporary directory will be created automatically when the service is started, restart the service in the local or test environment, delete the undertow.Port.* (if Tomcat, it is Tomcat.Port.*) directory generated under /tmp , and upload the file again.

4. Solution

1. Manually create the temporary directory (not recommended)

mkdir -p /tmp/undertow.8099.1610110889708131476/undertow1171565914461190366upload

PS: if the file is not uploaded for more than 10 days again, the same problem will occur. The symptoms are not the root cause.

2. Modify Linux system configuration (not recommended)

vim /usr/lib/tmpfiles.d/tmp.con
# Add at the end of the file, which means that the folder at the beginning of undertow will not be cleaned up
x /tmp/undertow*

PS: if multiple servers are deployed, each server needs to be modified.

3. Modify spring boot configuration file (recommended)

      # Specify a custom upload directory
      location: /mnt/tmp

PS: when using this method, you must ensure that /MNT/tmp exists. If it does not exist, the same error will occur. Therefore, it needs to be judged every time the service is started. If the directory exists, it will be ignored, and if it does not exist, it will be created. The code is as follows:

public class MultipartConfig {

    private String fileTempDir;

    MultipartConfigElement multipartConfigElement() {
        String os = System.getProperty("");
        // windows
            fileTempDir = "C:" + fileTempDir;
        }"fileTempDir:{}", fileTempDir);
        MultipartConfigFactory factory = new MultipartConfigFactory();
        File tmpDirFile = new File(fileTempDir);
        // Determine whether the folder exists
         if (!tmpDirFile.exists()) {
             //Create folder
            boolean mkdirSuccess = tmpDirFile.mkdirs();
  "create temp dir,result:{}", mkdirSuccess);
        return factory.createMultipartConfig();


[Solved] Lumen Error: Class redis does not exist

The company deployed a new project using lumen. The access interface reported a class redis does not exist error, which literally means that redis could not be found. After looking for a solution for a long time, it was later found that the redis plug-in was not added in composer.json, so the container could not find the redis service when loading. The solution is as follows:

1. Add “illuminate/redis” in composer.json: “^ 5.4”

2. Re execute composer install (if the lock file cannot be executed, delete composer.lock and then execute). After successful execution, the redis folder will be generated under vendor\illuminate

3. In bootstrap/app.php, add $app ->; register(Illuminate\Redis\RedisServiceProvider::class);

[Solved] Canal 1.1.5 Startup Error: caching_sha2_password Auth failed

1. Phenomenon caching_sha2_password Auth failed
        at ~[canal.parse.driver-1.1.5.jar:na]
        at ~[canal.parse.driver-1.1.5.jar:na]
        ... 4 common frames omitted
2021-11-20 16:43:40.852 [destination = example , address = / , EventParser] ERROR - destination:example[ connect / failure
2. Analysis and positioning

Since MySQL 8.0.3, the authentication plug-in uses caching by default_sha2_password

3. Solution

Solution: modify the authentication plug-in corresponding to the canal user to MySQL_native_password

mysql> select host,user,plugin from mysql.user ;
mysql> ALTER USER 'canal'@'%' IDENTIFIED WITH mysql_native_password BY 'password';


Euopenler 21.09 sudo Yum Update Error: Errors during downloading metadata for repository ‘EPOL’

sudo yum update error

Errors during downloading metadata for repository 'EPOL':
	-Status code: 404 for htpp://

Please note that:

sudo vi/etc/

Please note that


The correct address is:

baseurl= $basearch/

Save, exit and execute again.

[Solved] Python install kenlm error: ERROR: Command errored out with exit status 1: …

Use pip install kenlm error:

python/kenlm.cpp:6381:13: error: ‘PyThreadState {aka struct _ts}’ has no member named ‘exc_traceback’; did you mean ‘curexc_traceback’?
tstate->exc_traceback = *tb;
error: command ‘gcc’ failed with exit status 1

ERROR: Command errored out with exit status 1: …

It seems to be the GCC Issue

Use pypi-kenlm

pip install pypi-kenlm

SCP path contains special characters Error [How to Solve]


From a file on the local copy server, use the following command:

scp [email protected]:/home/test/files(202110~202111).xls .

Error reporting:
bash: – C: line 0: syntax error near unexpected token ` (‘


    1. enclose the entire path in single quotation marks;
  1. before the parentheses, add the escape character
scp [email protected]:'/home/test/files\(202110~202111\).xls' .

The semget function error: errno is set to 28 [How to Solve]

When running semget under Linux to create semaphores, it returns – 1 and the creation fails;

1. This function is a system function. You can only confirm the actual error code with errno, print errno through strError, and return no space left on device. Is the system space insufficient? Insufficient space to create semaphores?

2. Go to errno. H to check the error message enospc corresponding to the actual error code. What does this field mean?

3. Does the semget function have its own error field? Check the function manual: check the man Manual of semget function: a semaphore set has to be created but the system limit for the maximum number of semaphore sets (semmni), or the system-wide maximum number of semaphores. Semaphore exceeds system limit.

It is basically determined that it is caused by the system semaphore. First, temporarily modify the kernel semaphore parameters and run again to see whether it has been solved.

4. The following commands are used in viewing semaphores

#1)The sysctl command can view and set system kernel parameters
# The 4 corresponding values from left to right are SEMMSL, SEMMNS, SEMOPM and SEMMNI.
sysctl -a | grep sem #View the setting value of the system semaphore
kernel.sem = 250 32000 32 128

#2) There are three ways to modify: the numbers are for reference only
echo 610 86620 100 142 > /proc/sys/kernel/sem

sysctl -w kernel.sem="610 86620 100 142"

echo "kernel.sem=610 86620 100 142" >> /etc/sysctl.conf`

#3) View the current semaphore and pid of the system as well as user information, view more information and check --help
ipcs -s -p -c

#4) Delete the semaphore method of the specified semid, and check more usage --help
ipcrm -s semid

#5) Delete all semid semaphore methods
ipcrm  -asem

5. Here, in the process of finding semaphore resource leakage, in order to facilitate real-time viewing of semaphore information, the semaphore output is written into the script and printed circularly
echo “ipcs -s loop”

while [ 1 ]
	sleep 1
	ipcs -s

6. Note: the final problem here is to see why the semaphore in the code exceeds the limit. Normally, the semaphore will not exceed the system limit.

Spring integrated HBase error [How to Solve]

Problem 1
Replace the jar package with spring-data-hadoop-1.0.0.RELEASE version
Problem 2
Introduce hadoop-client-3.1.3.jar and hadoop-common-3.1.3.jar
Problem 3
java.lang.NoClassDefFoundError: org/apache/commons/configuration2/ConfigurationSolution
Introduce commons-configuration2-2.3.jar
Problem 4
java.lang.NoClassDefFoundError: org/apache/hadoop/util/PlatformName
Introduce hadoop-auth-3.1.3.jar
Problem 5
java.lang.NoClassDefFoundError: org/apache/hadoop/mapred/JobConf
Introduce hadoop-mapreduce-client-common-3.1.3.jar, hadoop-mapreduce-client-core-3.1.3.jar and
Problem 6
java.lang.NoClassDefFoundError: com/ctc/wstx/io/SystemId
Introduce woodstox-core-5.0.3.jar
Problem 7
java.lang.NoClassDefFoundError: com/google/common/collect/Interners
Introduce guava-30.1.1-jre.jar
Problem 8
java.lang.NoSuchMethodError:;)Lcom/google/ common/collect/MapMaker
Remove the google-collect-1.0.jar package, guava conflict
Problem 9
java.lang.NoClassDefFoundError: com/fasterxml/jackson/core/JsonGenerator
Introduce jackson-annotations-2.12.4.jar, jackson-core-2.12.4.jar and jackson-databind-2.12.4.jar
Problem 10
java.lang.NoClassDefFoundError: org/apache/hadoop/hbase/HBaseConfiguration
Introduce hbase-common-2.2.4.jar
Problem 11
java.lang.NoClassDefFoundError: org/apache/hadoop/hbase/client/HTableInterface
After searching for a long time, I found that it is written in the configuration file
<bean id=”htemplate” class=””>
<property name=”configuration” ref=”hbaseConfiguration”>
Comment it out Summary
Most of the problem is the lack of jar packages, Spring integration with Hbase requires 15 packages.
Among them.
These packages are also required when integrating HDFS

Android Studio compile error: build tools is corrupt [Solved]

Error Messages:
Installed Build Tools revision 32.0.0 rc1 is corrupted. Remove and install again using the SDK Manager.
Problem Cause.
File corruption in buildTools after AndroidSutdio version upgrade.

1 Modify d8.bat to dx.bat
2 Modify d8.jar to dx.jar

[Solved] Multithreading uses jsch to obtain a session for connection error: session.connect: connection reset


The project uses the spring batch framework. Multiple slices use jsch to obtain SFTP connections to read files and report errors

In fact, it is multithreading, using jsch to obtain the session connection and report an error

com.jcraft.jsch.JSchException: Session.connect: Connection reset

Jsch version



Various reasons have been found on the Internet. Some say the number of SSH terminal connections is limited, and some say there is a TCP connection problem. The final reason has not been found yet. Please inform us in the comment area


public static Session getSshSession(String sftpHost, int sftpPort, String userName, String password) {
	JSch jsch = new JSch();
	// GET sshSession
	Session sshSession = null;
	try {
		sshSession = jsch.getSession(userName, sftpHost, sftpPort);
	} catch (JSchException e) {
	if (StringUtils.isNotBlank(password)) {
	Properties sshConfig = new Properties();
	sshConfig.put("StrictHostKeyChecking", "no");
	return sshSession;

static void test() {
	for (int i = 1; i < 50; i++) {
		new Thread(() -> {
			Session sshSession = getSshSession("*.*.*.*", 22, "root", "***");
			try {
			} catch (Exception e) {
			} finally {



Create a channel pool using apache.commons.pool2

Since the SFTP configuration of the project is dynamic and not fixed, the following code is not encapsulated as a spring boot managed bean

Connection pool configuration:

public class ConnPoolConfig extends GenericObjectPoolConfig {
    public ConnPoolConfig() {
        // The minimum number of free objects in the object pool should be
        // The maximum capacity of the pool. The maximum number of objects to be stored in the pool
        // Check the validity of an object when it is borrowed from the pool.
        // How often the recycler thread performs idle object recovery (polling interval, in milliseconds)
        setTimeBetweenEvictionRunsMillis(60 * 60000);
        // Whether to verify the validity of the object when the recycler is scanning for idle objects.
        // If an object has not reached the specified threshold of idle time, and if testWhileIdle is configured to true
        // then it checks if the object is still valid, and if the object's resources have expired (e.g., the connection is disconnected), then he can be recycled.

Connection pool factory:

public class ConnPoolFactory extends BasePooledObjectFactory<ChannelSftp> {

    private String host;
    private Integer port;
    private String userName;
    private String password;
    private final String strictHostKeyChecking = "no";

    public ConnPoolFactory(String host, Integer port, String userName, String password) { = host;
        this.port = port;
        this.userName = userName;
        this.password = password;

    public ChannelSftp create() throws Exception {
        JSch jsch = new JSch();
        Session session = jsch.getSession(userName, host, port);
        Properties config = new Properties();
        config.put("StrictHostKeyChecking", strictHostKeyChecking);
        ChannelSftp channel = (ChannelSftp) session.openChannel("sftp");
        return channel;

    public PooledObject<ChannelSftp> wrap(ChannelSftp obj) {
        return new DefaultPooledObject<>(obj);

    // Destroy the object, if the object pool detects that an "object" idle timeout,
    // or if the operator detects that the "object" is no longer valid when "returning the object" to the object pool, then this will result in "object destruction";
    // The design of the "destroy object" operation is far different, but it must be clear:
    // When this method is called, the life of the "object" must end. If object is a thread, then the thread must exit at this point;
    // If object is a socket operation, then the socket must be closed;
    // If object is a file stream operation, then "data flush" is done and closed normally.
    public void destroyObject(PooledObject<ChannelSftp> pooledObject) throws Exception {
        Channel channel = pooledObject.getObject();
        Session session = channel.getSession();

    // Check if the object is "valid";
    // The Pool cannot hold invalid "objects", so the "background detection thread" will periodically check the validity of the "objects" in the Pool,
    // If the object is invalid, it will be removed from the Pool and destroyed;
    // In addition, when the caller gets an "object" from the Pool, it also checks the validity of the "object" to make sure that no "invalid" objects can be output to the caller;
    // When the caller returns the "object" to the Pool after use, the validity of the object is still checked. By validity,
    // The validity of the object is whether the object is in the expected state and can be used directly by the caller;
    // If the object is a socket, then its validity is whether the socket's channel is open/blocking timeout, etc.
    public boolean validateObject(PooledObject<ChannelSftp> pooledObject) {
        return pooledObject.getObject().isConnected();

    // "Activate" an object, an additional "activation" action when the Pool decides to remove an object for delivery to the caller,
    // For example, you can "reset" the list of parameters in the activateObject method to make it feel like a "newly created" object when the caller uses it;
    // If the object is a thread, you can reset the "thread break flag" in the "activate" operation, or wake up the thread from blocking, etc;
    // If the object is a socket, then you can refresh the channel in the "activate" operation,
    // or rebuild the link to the socket (if the socket is unexpectedly closed), etc.
    public void activateObject(PooledObject<ChannelSftp> pooledObject) throws Exception {
        ChannelSftp channelSftp = pooledObject.getObject();
        Session session = channelSftp.getSession();
        if (!session.isConnected()) {

    // "Passivate" the object, when the caller "returns the object", the Pool will "passivate the object".
    // The implication of passivate is that the "object" needs a "rest" for a while.
    // If the object is a socket, then you can passivateObject to clear the buffer and block the socket;
    // If the object is a thread, you can sleep the thread or wait for an object in the thread during the "passivate" operation.
    // Note that the methods activateObject and passivateObject need to correspond to each other to avoid deadlocks or confusion about the state of the "object".
    public void passivateObject(PooledObject<ChannelSftp> pooledObject) throws Exception {

Connection pool:

public class ConnPool extends GenericObjectPool<ChannelSftp> {

    private static final Map<String, ConnPool> MAP = new ConcurrentHashMap<>();

    private ConnPool(String host, Integer port, String userName, String password) {
        super(new ConnPoolFactory(host, port, userName, password), new ConnPoolConfig());

    public static ConnPool getConnPool(String host, Integer port, String userName, String password) {
        String key = host + ":" + port;
        ConnPool connPool = MAP.get(key);
        if (connPool == null) {
            synchronized (ConnPool.class) {
                connPool = MAP.get(key);
                if (connPool == null) {
                    connPool = new ConnPool(host, port, userName, password);
                    MAP.put(key, connPool);
        return connPool;

The connection pool supports the establishment of different pools for different remote IP

Tool class encapsulation:

public static ChannelSftp borrowChannel(ConnectionConfig connCfg) {
	ConnPool connPool = ConnPool.getConnPool(connCfg.getHost(), connCfg.getPort(), connCfg.getUserName(),
	try {
		return connPool.borrowObject();
	} catch (Exception e) {
		logger.error("Get channelSftp from pool fail", e);

public static void returnChannel(ConnectionConfig connCfg, ChannelSftp channel) {
	ConnPool connPool = ConnPool.getConnPool(connCfg.getHost(), connCfg.getPort(), connCfg.getUserName(),
	try {
	} catch (Exception e) {
		logger.error("Return channelSftp to pool fail", e);

No problem with the test:

static void test2() {
	AtomicInteger j = new AtomicInteger(0);
	for (int i = 0; i < 50; i++) {
		new Thread(() -> {
			ConnPool connPool = ConnPool.getConnPool("*", 22, "root", "*");
			System.out.println(connPool + "--" + j.getAndIncrement());
			ChannelSftp channelSftp = null;
			try {
				channelSftp = connPool.borrowObject();
			} catch (Exception e) {
			} finally {

How to Solve shiro Set sessionIdUrlRewritingEnabled Error (jessionid Removed)

Project scenario:

When using Shiro for authority authentication, the login address always carries the jeonid automatically for the first access. Now it needs to be removed and cannot be displayed.

Problem Description:

First, I searched Baidu and found that most solutions are to set SessionManager when defaultwebsecuritymanager is injected. This method also needs Shiro 1.3.2 or above. Coincidentally, my 1.3.0 is definitely not good. Go to POM directly to change the version number
annotation method:

    public DefaultWebSessionManager sessionManager(){
        DefaultWebSessionManager sessionManager = new DefaultWebSessionManager();
        return sessionManager;

    public DefaultWebSecurityManager securityManager(){
        DefaultWebSecurityManager securityManager = new DefaultWebSecurityManager();
        return securityManager;

XML mode:

		<bean id="sessionManager" class="org.apache.shiro.web.session.mgt.DefaultWebSessionManager">
		<property name="sessionIdUrlRewritingEnabled" value="false"/>

	<!-- Shiro Security Manager -->
	<bean id="securityManager" class="org.apache.shiro.web.mgt.DefaultWebSecurityManager">
		<property name="sessionManager" ref="sessionManager"/>

Remember to add package scanning in XML mode.My:

<context:component-scan base-package="config" />

Then run the project and report the same error as before. Anyway, there is no getter method.

Cause analysis:

During debugging, it is found that there are three versions of Shiro in the project, and two are from other modules, so it doesn’t matter. However, during debugging, it is found that the number of lines is not matched. After downloading the source code, it is found that this.sessionidurlrewritingenabled = true; There are no breakpoints in this line, and then the rebuild project, MVN clean install, repackaging and restarting idea are all used, and then the breakpoints can be interrupted. However, after debugging starts, the breakpoint icon turns into a circle slash, which is not the Shiro version of my current project at all
finally, you can only guess whether there are problems with multiple versions at the same time.


It is found that one of the three versions is very old 1.2.4, but there is no place to import the whole project. Then, according to the warehouse, it is still imported in the way of Shiro all, so the name is temporarily modified. After running the project, it can run as expected