[Bug] Fix correct min offset behavior in tiered storage

[lizhimins]

Before Creating the Bug Report

• I found a bug, not just asking a question, which should be created in GitHub Discussions.

• I have searched the GitHub Issues and GitHub Discussions of this repository and believe that this is not a duplicate.

• I have confirmed that this bug belongs to the current repository, not other repositories of RocketMQ.

Runtime platform environment

MacOS, JDK 11

RocketMQ version

branch: develop latest

JDK Version

JDK11

Describe the Bug

When resetting the consumer's offset to an earlier time to start consuming data from tiered storage, we use "clean expired" to remove older data because commitlog and consume queue file have their own lifecycles. However, it is possible for cq file to still exist while the commitlog no longer exists. In this scenario, the data pointed to by cq unit is empty and there is no reasonable "next begin offset" to provide, preventing the client from correcting the position and self-healing the consumption process.

Steps to Reproduce

No.

What Did You Expect to See?

Fix offset success.

What Did You See Instead?

No.

Additional Context

No response