JAVA异常的不正确处理方式

最近收到一个求助信息，异常堆栈如下：

```
java.lang.RuntimeException: FF1C1005
	at test_ssh.sftp.Pool.get(Pool.java:25) ~[test-ssh-0.0.1-SNAPSHOT.jar:na]
	... 116 common frames omitted
Caused by: java.lang.NullPointerException: null
	at test_ssh.sftp.Pool.get(Pool.java:23) ~[test-ssh-0.0.1-SNAPSHOT.jar:na]
```

从堆栈信息看就是一个NPE，其他什么也看不出，再看代码（简化的示例代码）：

 1 package test_ssh.sftp;
 2
 3 import java.util.concurrent.ConcurrentHashMap;
 4 import com.trilead.ssh2.Connection;
 5
 6 public class Pool {
 7     private static ConcurrentHashMap<String, Connection> pool = new ConcurrentHashMap<String, Connection>();
 8
 9     public synchronized static Connection get(String host, int ftpport) {
10         Connection connection = pool.get(host) == null ? connect(host, ftpport) : pool.get(host);
11         try {
12             connection.ping();
13         } catch (Exception e) {
14             throw new RuntimeException("FF1C1005", e);
15         }
16         return connection;
17
18     }
19
20     private static Connection connect(String host, int ftpport) {
21         Connection connection = new Connection(host, ftpport);
22         try {
23             connection.connect();
24         } catch (Throwable e) {
25             return null;
26         }
27         return connection;
28     }
29 }

代码没有在适当的点抛出异常，把原始异常信息丢了~先修正下代码，该抛的异常抛出来：

public synchronized static Connection get(String host, int ftpport) {
        Connection connection = pool.get(host) == null ? connect(host, ftpport) : pool.get(host);
        try {
            connection.ping();
        } catch (Exception e) {
        e.printStackTrace();
            throw new RuntimeException("FF1C1005", e);
        }
        return connection;
}

private static Connection connect(String host, int ftpport) {
        Connection connection = new Connection(host, ftpport);
        try {
            connection.connect();
        } catch (IOException e) {
            throw e;
        }
        return connection;

    }

}

重新打包运行，然后用新开会话运行tail -f 命令追踪日志，准备查看异常信息，然而并未如预期看到异常信息~难道IO被重定向了？ 停止服务，在控制终端下重新启动服务，果然日志刷新到标准输出了：

java.io.IOException: There was a problem while connecting to 10.246.186.37:10022
at com.trilead.ssh2.Connection.connect(Connection.java:843)
... 119 more
Caused by: java.io.IOException: Cannot negotiate, proposals do not match.
at com.trilead.ssh2.transport.KexManager.handleMessage(KexManager.java:399)

**Linux会话使用技巧**

服务端程序通常会以守护进程方式启动，不能随意重启，那如果想查看程序输出到stdout（通常不会这么做，但不排除意外）的信息怎么办呢？两个办法：

[1]tail -f /proc/PID/fd/1 这里1是标准输出，当然也可以是2 stderr
[2]sudo strace -p PID -e write