事务超时，发起方服务没有清理事务（clean transaction）的过程

- I have searched the [issues](https://github.com/codingapi/tx-lcn/issues) of this repository and believe that this is not a duplicate.

### 1. Bug Description
（不知道是不是bug，先请谅解）
服务A、B，A发起事务调用B。
测试A服务和B服务超时后事务的控制过程（通过断点实现，模拟业务处理时间过长造成整体事务超时），发现发起方A服务的事务没有清理（clean transaction）的过程，导致A服务锁表。参与方B服务有clean transaction的过程。

如下是参与方设置断点导致整体事务超时的事务过程：

如下是发起方设置断点导致整体事务超时的事务过程：

### 2. Environment:
- JDK version: 1.8
- OS: win7
- TX-LCN version: 5.0.1
- Others: Oracle 数据库

### 3. Exception Stacktrace

```
Paste your Exception Stacktrace here!
```

### 4. Tour Idea

Start by reproducing the timeout flow between services A and B using the described breakpoint-based delay, then compare the initiator and participant transaction cleanup sequences shown in the screenshots. Trace why the initiator retains its transaction and locks the table; done means the initiator performs cleanup after the overall transaction times out, without leaving the lock behind.

Written by the indexing model from the issue text.