redis updateGlobalTransactionDO unwatch need to be optimized

## Why you need it?
Is your feature request related to a problem? Please describe in details
redis模式里的updateGlobalTransactionDO在处理redis事务时使用了watch，但是使用unwatch的写法不统一，有些在抛出异常或者return之前执行了unwatch，有些则没有。需要搞清楚unwatch是不是必须的，然后保持写法一致

## How it could be?
A clear and concise description of what you want to happen. You can explain more about input of the feature, and output of it.
考虑到这个以下2件事

- watch+multi在redis服务端的原理是：watch命令对某个client操作的key打标记，在进行multi.exec命令时会检查同一client同一key是否有标记，有则拒绝整个exec
- jedis client目前的获取写法是twr写法，在抛出异常或者return时会被close掉，他的close实际上是returnResource给pool ，看源码可以看到里面会重置状态，发送unwatch命令

建议把返回之前的unwatch都去掉，不用引起误解。
如果我们觉得这样改动有风险，也可以统一把unwatch放在finally里面实现，因为update核心逻辑比较复杂，夹杂这个很难维护而且容易遗漏

## Other related information
Add any other context or screenshots about the feature request here.

Start at updateGlobalTransactionDO in the Redis mode and trace every Jedis watch, return, exception, and try-with-resources path. Read how returning a pooled Jedis instance resets state and sends UNWATCH, then determine a consistent cleanup policy and verify that Redis transaction behavior is unchanged across all paths.

Written by the indexing model from the issue text.