分布式环境下的串行化控制
项目背景：
分布式应用服务器，单点数据库

思路：
在数据库上做控制，保证事务串行化控制
给请求的资源分配一个唯一的锁，竞争到锁的线程开始执行业务
通过把请求锁的SQL写成SP，在代码执行的开始处调用

SP：
create or replace
PACKAGE BODY UTIL_LOCK IS

PROCEDURE request_lock(v_lockname in  VARCHAR2, p_retval OUT INTEGER, p_lockhandle OUT VARCHAR2) IS
  v_lockhandle LONG;
BEGIN
    DBMS_LOCK.allocate_unique(v_lockname, v_lockhandle);
    p_retval := dbms_lock.request(v_lockhandle,dbms_lock.x_mode,30000); 
    p_lockhandle :=v_lockhandle;
END request_lock;
------------------------
PROCEDURE release_lock(p_lockhandle IN VARCHAR2) IS
  p_out NUMBER;
BEGIN
    p_out := dbms_lock.release(p_lockhandle);
END release_lock;
------------------------
END UTIL_LOCK;

JAVA代码：
    private boolean lock(Connection conn, String resource)
    {
CallableStatement cs;
try
{
    String lStrSP = "{call UTIL_LOCK.request_lock(?,?,?)}";
    cs = conn.prepareCall(lStrSP);
    cs.setString(1, resource);
    cs.registerOutParameter(2, OracleTypes.NUMBER);
    cs.registerOutParameter(3, OracleTypes.VARCHAR);
    cs.execute();
    int ret = cs.getInt(2);
    this.handle = cs.getString(3);
    cs.close();
    return ret == 0;
}
catch (SQLException e)
{
    // TODO Auto-generated catch block
    e.printStackTrace();
    return false;
}
    }
调用：
if(lockutil.lock(conn,"resource")){
//place your code here
...
lockutil.release();
}