sql/logictest: add logictest for all expected 1PC txn statements

[nvanbenschoten]

First commit from #41324.

We expect a selection of simple single-statement mutations to hit the
"1-phase commit" transaction fast-path. #41320 demonstrated how critical
this is, as regressions here can cause major (> 50%) performance hits to
benchmarks and user workloads. This commit adds a logic test to validate
that these statements continue to hit this fast-path.

Release justification: Testing only.

[cockroach-teamcity]

This change is 

[RaduBerinde]

Have you looked at the tests in opt/exec/execbuilder/testdata/autocommit? I think these belong there

Reviewable status: complete! 0 of 0 LGTMs obtained (waiting on @jordanlewis)

[nvanbenschoten]

Yep, I based some of these tests off that file. #41320 is exactly why that level of testing alone isn't sufficient. In that issue, we saw that autocommit was set by the optimizer but that the deleteRangeNode simply ignored it. We need to test all the way down to the kv batch that is generated for these statements.

I'm happy to move these logic tests to that file though. I'm not sure what the difference between opt/exec/execbuilder/testdata and pkg/sql/logictest/testdata is at this point.

[RaduBerinde]

I would move them. The logic tests are meant to check that the statements do what they're supposed to. The execbuilder tests are for checking how we run them.

[nvanbenschoten]

I would move them. The logic tests are meant to check that the statements do what they're supposed to. The execbuilder tests are for checking how we run them.

Done. PTAL.

[RaduBerinde]

Reviewable status: complete! 1 of 0 LGTMs obtained (waiting on @jordanlewis)

[nvanbenschoten]

bors r+

[nvanbenschoten]

Bors was canceled before it ran.

bors r+

[nvanbenschoten]

bors crash:

{:timeout,
 {GenServer, :call,
  [
    BorsNG.GitHub,
    {:synthesize_commit,
     {{:installation, 117629}, 16563587},
     {%{
        branch: "staging",
        commit_message: "Merge #41358 #41371 #41372 #41379\n\n41358: sql/logictest: add logictest for all expected 1PC txn statements r=nvanbenschoten a=nvanbenschoten\n\nFirst commit from #41324.\r\n\r\nWe expect a selection of simple single-statement mutations to hit the\r\n\"1-phase commit\" transaction fast-path. #41320 demonstrated how critical\r\nthis is, as regressions here can cause major (> 50%) performance hits to\r\nbenchmarks and user workloads. This commit adds a logic test to validate\r\nthat these statements continue to hit this fast-path.\r\n\r\nRelease justification: Testing only.\n\n41371: kv: avoid allocations in client.Txn constructor r=nvanbenschoten a=nvanbenschoten\n\nThis PR contains two small wins that help speed up the client.Txn constructor, which is responsible for **2.90%** of a CPU profile when running Sysbench's `oltp_point_select` workload. The biggest win here will come from addressing https://github.com/cockroachdb/cockroach/issues/32508.\r\n\r\n#### kv: lazily create *RangeIterator in txnPipeliner\r\n\r\nThis allocation was responsible for **0.34%** of a CPU profile when running Sysbench's `oltp_point_select` workload.\r\n\r\n#### kv: only re-alloc refresh spans in augmentMetaLocked if necessary\r\n\r\nThis was responsible for **0.057%** of a CPU profile when running Sysbench's `oltp_point_select` workload.\r\n\r\nRelease justification: None. These can wait for the branch to split.\n\n41372: sql/pgwire: statically allocate format code slice for all text args/cols r=nvanbenschoten a=nvanbenschoten\n\nThis allocation was responsible for **0.8%** of a CPU profile when running\r\nSysbench's oltp_point_select workload.\r\n\r\nRelease justification: Probably none, although this does look very safe.\r\n\r\nRelease note: None\n\n41379: pkg/sql: use ring.Buffer in StmtBuf r=nvanbenschoten a=nvanbenschoten\n\nThe StmtBuf has the exact access patterns typically associated with a ring buffer. Command instances are added to the back of the buffer and trimmed from the front of the buffer. These operations are often performed in lockstep, but that is not always the case, so we need the buffer to be able to grow.\r\n\r\n`ring.Buffer` provides exactly these semantics and it allows us to avoid memory allocations on each Command in `StmtBuf.Push`.\r\n\r\nRelease justification: None. Don't merge until branch is split.\n\nCo-authored-by: Nathan VanBenschoten <nvanbenschoten@gmail.com>\n",
        committer: %{
          email: "bors@cockroachlabs.com",
          name: "craig[bot]"
        },
        parents: ["6d79c7a4f36acd10c410a87ab9bddf3580f28748",
         "a0dde23104734fc2f0ae88397896686433963630",
         "9d682e9c6c112f9246456a1548dc4d3842031257",
         "fdbf14e2395444e68535a30d5be3145750e5b296",
         "42ec21dd7deef1bd5616ea663db0d6b4fe436c91"],
        tree: "6291963028f8f3aeadff552af89a617a6ee464ab"
      }}},
    10000
  ]}}

bors r+

[craig]

Build succeeded

• GitHub CI (Cockroach)