release-24.3: sql/row: fix multi-range reads from PCR standby

[blathers-crl]

Backport 1/1 commits from #132854 and 1/1 commits from #133589 on behalf of @michae2.

/cc @cockroachdb/release

---

sql/row: fix multi-range reads from PCR standby

Reads from tables with external row data (i.e. reads from a PCR standby cluster) need to use the fixed timestamp specified by the external row data. This timestamp might be different from the transaction timestamp, so we were explicitly setting BatchRequest.Timestamp in kv_batch_fetcher.

The KV API only allows BatchRequest.Timestamp to be set for non-transactional requests (i.e. requests sent with a NonTransactionalSender, which is a CrossRangeTxnWrapperSender in this case). We were using a NonTransactionalSender, but this had two problems:

1. CrossRangeTxnWrapperSender in turn sends the BatchRequest with a transactional sender, which again does not allow BatchRequest.Timestamp to be set.
2. CrossRangeTxnWrapperSender uses kv.(*Txn).CommitInBatch, which does not provide the 1-to-1 request-response guarantee required by txnKVFetcher. It is kv.(*Txn).Send which provides this guarantee.

Because of these two problems, whenever the txnKVFetcher would send a multi-range-spanning BatchRequest to CrossRangeTxnWrapperSender, it would either fail with a "transactional request must not set batch timestamp" error or would return an unexpected number of responses, violating the txnKVFetcher's assumed mapping from request to response.

To fix both these problems, instead of using a NonTransactionalSender, change the txnKVFetcher to open a new root transaction with the correct fixed timestamp, and then use txn.Send.

Fixes: #132608

Release note: None

---

sqlccl: skip TestStandbyRead under duress

Fixes: #133243

Release note: None

---

Release justification: fix for a GA-blocker bug.

[blathers-crl]

Thanks for opening a backport.

Please check the backport criteria before merging:

• Backports should only be created for serious
  issues or test-only changes.
• Backports should not break backwards-compatibility.
• Backports should change as little code as possible.
• Backports should not change on-disk formats or node communication protocols.
• Backports should not add new functionality (except as defined
  here).
• Backports must not add, edit, or otherwise modify cluster versions; or add version gates.
• All backports must be reviewed by the owning areas TL. For more information as to how that review should be conducted, please consult the backport
  policy.

If your backport adds new functionality, please ensure that the following additional criteria are satisfied:

• There is a high priority need for the functionality that cannot wait until the next release and is difficult to address in another way.
• The new functionality is additive-only and only runs for clusters which have specifically “opted in” to it (e.g. by a cluster setting).
• New code is protected by a conditional check that is trivial to verify and ensures that it only runs for opt-in clusters. State changes must be further protected such that nodes running old binaries will not be negatively impacted by the new state (with a mixed version test added).
• The PM and TL on the team that owns the changed code have signed off that the change obeys the above rules.
• Your backport must be accompanied by a post to the appropriate Slack
  channel (#db-backports-point-releases or #db-backports-XX-X-release) for awareness and discussion.

Also, please add a brief release justification to the body of your PR to justify this
backport.

[cockroach-teamcity]

This change is 

[michae2]

Failure looks like #132638, which I think is unrelated.

[michae2]

Note this will also need #133589.

[DrewKimball]

Reviewed 3 of 3 files at r1, all commit messages.
Reviewable status: complete! 0 of 0 LGTMs obtained (waiting on @dt, @mgartner, @michae2, and @nvanbenschoten)