release-24.3: import: do not revert data if cancel called after success

[blathers-crl]

Backport 1/1 commits from #159627 on behalf of @jeffswenson.
Backport 1/1 commits from #161475 on behalf of @ZhouXing19.

---

Currently, it is possible that canceling an import after the job marked the descriptor as online could cause the non-mvcc compliant rollback code to run on an online range. Note, this fix is still slightly incomplete. If there is a zombie import job revert logic, that zombie may revert changes its not supposed to, but the liveness subsystem makes zombie jobs unlikely.

Release note: Fixes an unlikely bug where import rollback could revert writes in an online descriptor. This is not a silent failure. It would only occur if the job entered a failed or cancelled state after the actual import succeeded and marked the descriptor as on line.
Fixes: #159603

---

Release justification: bug fix.

[blathers-crl]

Thanks for opening a backport.

Before merging, please confirm that it falls into one of the following categories (select one):

• Non-production code changes. Includes test-only changes, build system changes, etc.
• Fixes for serious issues. Defined in the policy as correctness, stability, or security issues, data corruption/loss, significant performance regressions, breaking working and widely used functionality, or an inability to detect and debug production issues.
• Other approved changes. These changes must be gated behind a disabled-by-default feature flag unless there is a strong justification not to.

Add a brief release justification to the PR description explaining your selection.

Also, confirm that the change does not break backward compatibility and complies with all aspects of the backport policy.

All backports must be reviewed by the TL and EM for the owning area.

[cockroach-teamcity]

This change is 

[blathers-crl]

❌ PR #159900 does not comply with backport policy

Confidence: high
Explanation: The pull request modifies production code in 'pkg/sql/importer/import_job.go' by changing the handling logic for import job cancellations. Specifically, it prevents reverting data if a cancel is called after an import job has been successfully completed and the table descriptor has gone online. This change addresses a significant issue where rollback could incorrectly affect online data. Although this qualifies as fixing a stability or data corruption issue (meeting the critical bug criteria), there is no evidence provided in the PR that the changes are guarded by a feature flag. Furthermore, the release justification only mentions this as a 'bug fix' without qualifying it as addressing a critical bug.
Recommendation: Gate the PR behind a feature flag or provide additional justification that explains why this fix qualifies as a critical bug and can be safely backported without a feature flag

_{🦉 Hoot! I am a Blathers, a bot for CockroachDB. My owner is dev-inf.}

[michae2]

LGTM

[github-actions]

Reminder: it has been 2 weeks please merge or close your backport!