项目中有2个定时任务要执行，并且调用相同的公共类查询hbase

public class CronA {

  @Autowired
  GetDataByAccounts a;

  @Async
  @Scheduled(fixedRate = 1000*60*1)
   public void A(){
       String key = "xx";
      a.getDataByAccounts(key);
  }
}

public class CronB {
  @Autowired
  GetDataByAccounts b;

  @Async
  @Scheduled(fixedRate = 1000*60*3)
   public void B(){
       String key = "xx";
       b.getDataByAccounts(key);
  }
}

公共类的代码如下：

public class GetDataByAccounts{

  Connection conn;
  Configuration conf;

  public String getDataByAccounts(String key){

       // Connection conn = null;
       // Configuration conf = HBaseConfiguration.create();
       // hbase操作省略

   }

}

上线后，报错：
error日志：
2021-02-07 15:27:18 [task-2] ERROR o.a.h.h.c.AsyncRequestFutureImpl -- - Cannot get replica 0 location for {"cacheBlocks":true,"totalColumns":0,"row":"xx","families":{},"maxVersions":1,"timeRange":[0,9223372036854775807]}
2021-02-07 15:27:18 [task-2] ERROR c.i.tools.service.GetDataByAccounts -- - 操作hbase异常:Failed 33 actions: hconnection-0xd440f89 closed: 33 times, servers with issues: null
org.apache.hadoop.hbase.client.RetriesExhaustedWithDetailsException: Failed 33 actions: hconnection-0xd440f89 closed: 33 times, servers with issues: null
warn日志：
2021-02-07 15:27:18 [task-2] WARN o.a.h.h.c.AsyncRequestFutureImpl -- - id=1, task rejected by pool. Unexpected. Server=n2,16020,1609750056695
java.util.concurrent.RejectedExecutionException: Task java.util.concurrent.FutureTask@7803f2e2 rejected from java.util.concurrent.ThreadPoolExecutor@4cdfbc3a[Terminated, pool size = 0, active threads = 0, queued tasks = 0, completed tasks = 14]
查阅发现：可能的原因是table被提前关闭了

结合代码来看：
1、使用@Autowired注解的对象a,b，打印出地址：发现是同一个对象，从而共享了同一个Connection和Table，构造函数也只调用了1次
2、2个定时任务第一次运行的时间相同，导致其中1个操作结束后关闭了table，另一个操作报错，并且是一定概率发生

解决方法是：
将Connection和Configuration的定义由类变量改为函数内的局部变量，即：GetDataByAccounts中注释的2行代码