[FEATURE] [tez] Avoid recompute succeeded task.

### Code of Conduct

- [X] I agree to follow this project's [Code of Conduct](https://www.apache.org/foundation/policies/conduct)

### Search before asking

- [X] I have searched in the [issues](https://github.com/apache/incubator-uniffle/issues?q=is%3Aissue) and found no similar issues.

### Describe the feature

In general local shuffle, some succeeded task may be recomputed. For example, NM is unhealthy, reduce can not read the shuffle date from map which have run successfully in unhealthy NM.
But for remote shuffle, the shuffle data have store in remote shuffle service, recompute the succeeded task which have run in unhealthy node is unnecessary.

### Motivation

_No response_

### Describe the solution

_No response_

### Additional context

_No response_

### Are you willing to submit PR?

- [X] Yes I am willing to submit a PR!

The issue names Tez integration and remote shuffle handling but no files or tests. Start by tracing how succeeded tasks are recomputed for local versus remote shuffle when a node is unhealthy. Done means a succeeded task whose shuffle data is stored remotely is not recomputed, with regression coverage for that failure scenario.

Written by the indexing model from the issue text.