HBase学习 - 深入理解（3）

谁践踏了优雅 2022-03-07 02:11 258阅读 0赞

**目录：**

1.  HBase与Schema设计
    
    1.  Schema 创建
    2.  column families的数量：
    3.  Rowkey 设计：Hbase 在进行表设计过程中如何进行列族和 RowKey 的设计？  
        尽量在你的应用中使用一个列族。尽量最小化行和列的大小
    4.  版本数量：Hbase 有没有并发问题？Hbase 如何实现自己的 MVVC 的？  
        版本数量可以设置
    5.  支持的数据类型
    6.  Joins
    7.  生存时间 (TTL)
    8.  保留删除的单元
    9.  第二索引和替代查询路径
    10. 限制
    11. 模式设计用例
    12. 操作和性能配置选项
2.  HBase与MapReduce设计
3.  HBase安全  
    Hbase 集群中 HRegionServer 宕机如何解决？  
    HRegionServer故障转移的规则：不是从空闲的HRegionServer中选取（不用zookeeper进行选举），而是根据集群HRegionServer，从顺时针方向下一个中选取，也就不需要实时监控RegionServer，
4.  性能调优  
    Hbase 的数据热点问题发现和解决办法：写流程  
    提高 Hbase 的读写性能的通用做法
5.  故障排除和调试HBase
6.  Zookeeper
7.  HBase处理数据：读、写、更新、删除流程的简单理解
8.  Java 操作过滤器 - filter的使用：扫描、计数和截断

### 一.HBase处理数据：读、写、更新、删除流程的简单理解 ###

**1.****写流程**

1.  Client先访问zookeeper，zookeeper读取meta表的region位置，然后读取meta表中的数据。meta表中又存储了用户的region信息。
2.  client 将数据分组：将数据记录先按照region分组，然后再按照RegionServer分组
3.  client 往RegionServer发送写数据请求
4.  往region中写数据![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70][]
    
    1.  获取Region操作锁（读写锁）
    2.  依次获取各行行锁
    3.  写入到memstore中
    4.  释放已获取的行锁
    5.  写数据到WAL中
    6.  释放Region锁
5.  写完之后flush
6.  HFile合并 - Compaction
    
    1.  小合并
    2.  大合并
7.  如果在数据插入的时候region超过了设置的大小，就会split。

**详细步骤：**  
![20190830181614987.png][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 1][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 2][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 3][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 4][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 5][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 6][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 7][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 8][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 9][]

Compaction（压平）说明：  
![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 10][]  
![20190511144537504.png][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 11][]

split说明：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 12][]

**说明：**

hbase表的结构：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 13][]

### put数据的流程图说明： ###

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 14][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 15][]

### 写数据总体流程： ###

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 16][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 17][]

**2.读流程**

**HBase读数据流程**

1.  Client先访问zookeeper，从meta表读取region的位置，然后读取meta表中的数据。meta中又存储了用户表的region信息。
2.  根据namespace、表名和rowkey在meta表中找到对应的region信息
3.  找到这个region对应的regionserver
4.  查找对应的region
5.  先从MemStore找数据，如果没有，再到StoreFile上读(为了读取的效率)。

**详细流程：**

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 18][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 19][]

1.  一个列族对应一个Scanner.注意下图中的列族名。  
      
    ![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 20][]
2.  Next说明：  
    ![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 21][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 22][]

filter说明：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 23][]

![2019051118234738.png][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 24][]

key-value说明：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 25][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 26][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 27][]

key-value结构说明：、

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 28][]

## ##

### 二.Java操作过滤器 ###

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 29][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 30][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 31][]

参考博客：[https://blog.csdn.net/u010967382/article/details/37653177][https_blog.csdn.net_u010967382_article_details_37653177]

普通过滤器

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 32][]

多条件过滤器

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 33][]

**练习：**

1.  创建项目
2.  添加依赖
    
        <dependency>
            	<groupId>org.apache.hbase</groupId>
            	<artifactId>hbase-client</artifactId>
            	<version>1.2.6</version>
            </dependency>
3.  编写代码  
    第一步：先建表  
    ![2019051312153627.png][]  
    第二步：插入数据
    
        public class Test {
        	public static void main(String[] args) throws IOException {
        		Configuration conf=new Configuration();
        		conf.set("hbase.zookeeper.property.clientPort","2181");
        		conf.set("hbase.zookeeper.quorum", "192.168.200.10");
        		//如果是zookeeper集群
        		//conf.set("hbase.zookeeper.quorum","192.168.200.10,192.168.200.11,192.168.200.12");
        		Connection conn=ConnectionFactory.createConnection(conf);
        		//先插入数据
        		putListData(conn);
        		conn.close();
        	}
        	
        	/*向hbase插入多条值*/
        	private static void putListData(Connection conn) throws IOException {
        		//先找到表
        		Table t01=conn.getTable(TableName.valueOf("t01"));
        		List<Put> list=new ArrayList();
        		Put put=null;
        		for(int i=0;i<100;i++) {
        			put =new Put(Bytes.toBytes("rowkey001"+i));
        			put.addColumn("cf01".getBytes(), Bytes.toBytes("name"), Bytes.toBytes("javazs"+i));
        			list.add(put);
        		}
        		t01.put(list);
        		t01.close();
        
        	}
        }
    
    第三步：编写过滤器  
    1.第一种：普通过滤器的使用：name为 "javazs9" 开头的数据
    
        public class Test {
        	public static void main(String[] args) throws IOException {
        		Configuration conf=new Configuration();
        		conf.set("hbase.zookeeper.property.clientPort","2181");
        		conf.set("hbase.zookeeper.quorum", "192.168.200.10");
        		//如果是zookeeper集群
        		//conf.set("hbase.zookeeper.quorum","192.168.200.10,192.168.200.11,192.168.200.12");
        		Connection conn=ConnectionFactory.createConnection(conf);
        		//先插入数据
        		//putListData(conn);
        		
        		//过滤器的使用
        		Table t01=conn.getTable(TableName.valueOf("t01"));
        		Scan scan=new Scan();
        		RegexStringComparator comparator=new RegexStringComparator("javazs9");
        		
        		/*第一个参数是列族
        		第二个参数是字段名
        		第三个参数是操作关系大于小于等等
        		第四个参数是比较规则*/
        		Filter filter=new SingleColumnValueFilter(Bytes.toBytes("cf01"), "name".getBytes(), CompareFilter.CompareOp.EQUAL, comparator);
        		//将filter添加到scan中
        		scan.setFilter(filter);
        		//每次从hbase中取出10条数据，以批量的方式取出
        		scan.setCaching(10);
        		
        		// Reading values from scan result
        	        for (Result result = scanner.next(); result != null; result = scanner.next()) {
        	    	    byte [] value = result.getValue(Bytes.toBytes("cf01"), Bytes.toBytes("name"));
        	    	    String name = Bytes.toString(value);
        	    	    //System.out.println("Found row : " + result);
        	    	    System.out.println("name="+name);
        	        }
        
        		conn.close();
        	}
        }
    
    结果：  
    ![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 34][]  
      
    2.第二种：如果列名不存在的情况，怎么实现过滤  
    ![20190516152254342.png][]
    
        public class Test {
        	public static void main(String[] args) throws IOException {
        		Configuration conf=new Configuration();
        		conf.set("hbase.zookeeper.property.clientPort","2181");
        		conf.set("hbase.zookeeper.quorum", "192.168.200.10");
        		//如果是zookeeper集群
        		//conf.set("hbase.zookeeper.quorum","192.168.200.10,192.168.200.11,192.168.200.12");
        		Connection conn=ConnectionFactory.createConnection(conf);
        		//先插入数据
        		//putListData(conn);
        		
        		//过滤器的使用
        		Table t01=conn.getTable(TableName.valueOf("t01"));
        		Scan scan=new Scan();
        		//RegexStringComparator comparator=new RegexStringComparator("18");
        		
        		/*第一个参数是列族
        		第二个参数是字段名
        		第三个参数是操作关系大于小于等等
        		第四个参数是比较规则*/
        		//在默认情况下，hbase在使用过滤器的时候，当某一行没有该字段时，hbase采用的方式为默认匹配：即默认为符合要求
        		Filter filter=new SingleColumnValueFilter(Bytes.toBytes("cf01"), "age".getBytes(), CompareFilter.CompareOp.EQUAL, "18".getBytes());
        		//将filter添加到scan中
        		scan.setFilter(filter);
        		//每次从hbase中取出10条数据，以批量的方式取出
        		scan.setCaching(10);
        		
        		// Getting the scan result
        	        ResultScanner scanner = t01.getScanner(scan);
        
        	        // Reading values from scan result
        	        for (Result result = scanner.next(); result != null; result = scanner.next()) {
        	    	    byte [] value = result.getValue(Bytes.toBytes("cf01"), Bytes.toBytes("name"));
        	    	    String name = Bytes.toString(value);
        	    	    //System.out.println("Found row : " + result);
        	    	    System.out.println("name="+name);
        	        }
        
        		conn.close();
        	}
        }
    
    结果：返回是所有数据，显然不符合我们的要求  
    改进：  
     
    
        //在默认情况下，hbase在使用过滤器的时候，当某一行没有该字段时，hbase采用的方式为默认匹配：即默认为符合要求
        SingleColumnValueFilter filter=new SingleColumnValueFilter(Bytes.toBytes("cf01"), "age".getBytes(), CompareFilter.CompareOp.EQUAL, "18".getBytes());
        filter.setFilterIfMissing(true);//添加此行
    
    结果：  
    ![20190516153314709.png][]  
      
    3.第三种：多个过滤器的使用：年龄大于18，value值以zs98开头的值  
    第一步：更新rowkey00198的值  
    ![20190516153500483.png][]  
    第二步：代码
    
        public class Test {
        	public static void main(String[] args) throws IOException {
        		Configuration conf=new Configuration();
        		conf.set("hbase.zookeeper.property.clientPort","2181");
        		conf.set("hbase.zookeeper.quorum", "192.168.200.10");
        		//如果是zookeeper集群
        		//conf.set("hbase.zookeeper.quorum","192.168.200.10,192.168.200.11,192.168.200.12");
        		Connection conn=ConnectionFactory.createConnection(conf);
        		//先插入数据
        		//putListData(conn);
        		
        		//过滤器的使用
        		Table t01=conn.getTable(TableName.valueOf("t01"));
        		Scan scan=new Scan();
        		FilterList filterList=new FilterList();
        		
        		//第一个Filter
        		RegexStringComparator comparator=new RegexStringComparator("javazs9");
        		/*第一个参数是列族
        		第二个参数是字段名
        		第三个参数是操作关系大于小于等等
        		第四个参数是比较规则*/
        		SingleColumnValueFilter filter1=new SingleColumnValueFilter(Bytes.toBytes("cf01"), "name".getBytes(), CompareFilter.CompareOp.EQUAL,comparator);
        		
        		//第二个Filter
        		//在默认情况下，hbase在使用过滤器的时候，当某一行没有该字段时，hbase采用的方式为默认匹配：即默认为符合要求
        		SingleColumnValueFilter filter2=new SingleColumnValueFilter(Bytes.toBytes("cf01"), "age".getBytes(), CompareFilter.CompareOp.GREATER_OR_EQUAL,"18".getBytes());
        		//如果该行没有age这个字段将不会匹配
        		//在什么条件下不需要设置这个属性？---在表中的所有行数据都有你需要查找的字段名时，可以不用设置
        		filter2.setFilterIfMissing(true);
        		filterList.addFilter(filter1);
        		filterList.addFilter(filter2);
        		//将filter添加到scan中
        		scan.setFilter(filterList);
        		//每次从hbase中取出10条数据，以批量的方式取出
        		scan.setCaching(10);
        		
        		// Getting the scan result
        	        ResultScanner scanner = t01.getScanner(scan);
        
        	        // Reading values from scan result
        	        for (Result result = scanner.next(); result != null; result = scanner.next()) {
        	    	    byte [] value = result.getValue(Bytes.toBytes("cf01"), Bytes.toBytes("name"));
        	    	    String name = Bytes.toString(value);
        	    	    //System.out.println("Found row : " + result);
        	    	    System.out.println("name="+name);
        	        }
        
        		conn.close();
        	}
        }
4.  运行  
    ![20190516161219601.png][]

[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70]: /images/20220307/997e2d8e06bd4afeac5736d7dedf57a4.png
[20190830181614987.png]: /images/20220307/0e32df5cbee1415096d9b11507b9ec2b.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 1]: /images/20220307/2fb62cb2ed4540a4a4fcd410362c0c56.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 2]: /images/20220307/d218882d506444348ea2f40644bfe30a.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 3]: /images/20220307/f252f8c814f1405a8912c7f424e90cc9.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 4]: /images/20220307/b2cb56be39064ee6b0e3e3a54f407079.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 5]: /images/20220307/0058d070288e4eb5830fdf4b01ee2472.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 6]: /images/20220307/6902d64c5065429ebf4705c35244d59a.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 7]: /images/20220307/7f0d5fc177d94e9099baa094f5e2e7e4.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 8]: /images/20220307/7415e74ad4bf43a3a3164ccc1febbbd6.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 9]: /images/20220307/b8275a11f1884d03a84a3bfc7f240e62.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 10]: /images/20220307/6f4ecc5a0a1548f2bf7f2db64773b46e.png
[20190511144537504.png]: /images/20220307/1faecaf4fe584053bd98969c6570386f.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 11]: /images/20220307/3c0f672057794feebfe1406b5b29b52a.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 12]: /images/20220307/686ba8b033cd4d7f9d2bcc4a481d3578.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 13]: /images/20220307/8e089d4d070a4fb98e394d6a20db45cc.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 14]: /images/20220307/799052b345964bf1801e8ac4e2d4e395.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 15]: /images/20220307/f8f300771b764b499985425603d62f24.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 16]: /images/20220307/f226c9683bbf4331a044eacd6a02c4b3.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 17]: /images/20220307/2a0512dcc08f4e86a8a8e5fc876c1f32.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 18]: /images/20220307/0ab1c5b0efbf43649d054e30b4176797.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 19]: /images/20220307/bb501ec84fbc4f03acc2e781809f55bc.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 20]: /images/20220307/94829205c5614098874ac44094a277f3.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 21]: /images/20220307/76db27e73e9249d6a8a37e20e6c26381.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 22]: /images/20220307/28adf7cc34cc4f6097351b7083276a1e.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 23]: /images/20220307/9ad515fd9b90467385b780de287c2e82.png
[2019051118234738.png]: /images/20220307/724b4f70af0b435fa454e8a5361ba9bc.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 24]: /images/20220307/35087ad9550548c3bd9fa816c2924277.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 25]: /images/20220307/1a6671a5742540149fc691209159ecce.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 26]: /images/20220307/8427e46db39a4ddf9440487988b57614.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 27]: /images/20220307/39bb9778927e43d2b89f4d2357853cf5.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 28]: /images/20220307/893ed60b187b438ba65b1a658f8b1c41.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 29]: /images/20220307/7ac871cb003e405f88b33cf14dba6201.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 30]: /images/20220307/942a09fd20cd4ec0a0d15c11069efc14.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 31]: /images/20220307/bf4404e84b3d4dc2b3b2c03972f1a74f.png
[https_blog.csdn.net_u010967382_article_details_37653177]: https://blog.csdn.net/u010967382/article/details/37653177
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 32]: /images/20220307/4e46fdc056aa4bab83fa514c6fa54569.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 33]: /images/20220307/2ef7695ee11642dc8140b4401416ea8f.png
[2019051312153627.png]: /images/20220307/ab06f720d3394cf6bd3487cb15bf7876.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lpZ3VhbmdfODIw_size_16_color_FFFFFF_t_70 34]: /images/20220307/668067d8a47b4540a00a61ee2a44f807.png
[20190516152254342.png]: /images/20220307/3eee288465914a53aa6bec8603bfc347.png
[20190516153314709.png]: /images/20220307/fa77ce0bbb634544a86bea208849eb61.png
[20190516153500483.png]: /images/20220307/e590631f5bc44e388f2ef4cfaf1a6149.png
[20190516161219601.png]: /images/20220307/87a40ff7da1e4352b9921c112b498f64.png