fix(p2p-11): harden reader and listener lifecycle (#671)

[jlon]

Summary

• Harden reader failure handling and data-plane listener lifecycle.
• Make this the first redirect slice that is reasonable for rollout when stacked on p2p-10.

Design

• Keep worker fallback correct and prevent broken remote readers from re-entering the pool.
• Isolate listener retry, shutdown, and generation safety from later performance work.

Implementation

• Fix last-replica failure handling, remote reader cleanup, and block client pool hygiene.
• Add listener retry/stop/start hardening and the FsContext and pool lifetime fixes needed for P2P lifecycle.

Verification

• Verify broken remote readers are not returned to the pool.
• Verify stop/start is clean and transient listener failures self-recover.

[jlon]

Part 11/17 of the split P2P series that supersedes #671.

This PR targets , so the visible diff is cumulative. Review the slice described in this PR body instead of treating the whole diff as new.

Intended review delta is p2p-11 on top of #753.

[jlon]

Review note: this is part 11/17 of the split P2P series that supersedes #671.

This PR targets main, so the visible diff is cumulative. Please review the slice described in this PR body.

Review after #753.