release-22.2: storage: introduce test-build assertion iterator

[jbowens]

Backport #90859, #96685 and #97222 to mangle the buffers returned by Pebble
iterators to ensure MVCC and Cockroach code respects the Pebble iterator memory
lifetimes.

The RangeBounds() and RangeKeys() validity assertions are omitted, since 22.2
allowed these methods to be invoked on an invalid iterator or an iterator
positioned at a point key.

Epic: None
Release note: None
Release justification: Non-production code changes and fix to high-severity MVCC iterator bug.

[blathers-crl]

Thanks for opening a backport.

Please check the backport criteria before merging:

• Patches should only be created for serious issues or test-only changes.
• Patches should not break backwards-compatibility.
• Patches should change as little code as possible.
• Patches should not change on-disk formats or node communication protocols.
• Patches should not add new functionality.
• Patches must not add, edit, or otherwise modify cluster versions; or add version gates.

If some of the basic criteria cannot be satisfied, ensure that the exceptional criteria are satisfied within.

• 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.
• The PM and TL on the team that owns the changed code have signed off that the change obeys the above rules.

Add a brief release justification to the body of your PR to justify this backport.

Some other things to consider:

• What did we do to ensure that a user that doesn’t know & care about this backport, has no idea that it happened?
• Will this work in a cluster of mixed patch versions? Did we test that?
• If a user upgrades a patch version, uses this feature, and then downgrades, what happens?

[cockroach-teamcity]

This change is 

[erikgrinaker]

Thanks for catching this, pointSynthesizingIter changes LGTM. Mind running a quick MVCCGet benchmark to see the impact?

Reviewed 19 of 19 files at r1, 1 of 1 files at r2, 1 of 1 files at r3, all commit messages.
Reviewable status: complete! 0 of 0 LGTMs obtained (waiting on @jbowens and @RaduBerinde)

---

pkg/storage/sst.go line 615 at r1 (raw file):

				// Re-retrieve the buffers.
				if extHasRange {
					extRangeKeys = extIter.RangeKeys()

Consider pulling this out to a separate commit for visibility.

---

pkg/storage/pebbleiter/crdb_test_off.go line 19 at r1 (raw file):

// Iterator redefines *pebble.Iterator.
type Iterator = *pebble.Iterator

Nice trick to avoid dynamic dispatch here.

[jbowens]

TFTR!

Mind running a quick MVCCGet benchmark to see the impact?

For sure. running them on my gceworker now

Reviewable status: complete! 0 of 0 LGTMs obtained (waiting on @erikgrinaker and @RaduBerinde)

---

pkg/storage/sst.go line 615 at r1 (raw file):

Previously, erikgrinaker (Erik Grinaker) wrote…

Consider pulling this out to a separate commit for visibility.

Good call, done.

[jbowens]

Impact looks not horrible:

name                                                                     old time/op    new time/op    delta
MVCCGet_Pebble/batch=false/versions=1/valueSize=8/numRangeKeys=0-24        6.29µs ± 1%    6.32µs ± 1%    ~     (p=0.310 n=5+5)
MVCCGet_Pebble/batch=false/versions=1/valueSize=8/numRangeKeys=1-24        11.5µs ± 1%    11.7µs ± 1%    ~     (p=0.056 n=5+5)
MVCCGet_Pebble/batch=false/versions=1/valueSize=8/numRangeKeys=100-24       133µs ± 1%     129µs ± 1%  -3.01%  (p=0.008 n=5+5)
MVCCGet_Pebble/batch=false/versions=10/valueSize=8/numRangeKeys=0-24       25.8µs ± 0%    25.7µs ± 1%    ~     (p=0.421 n=5+5)
MVCCGet_Pebble/batch=false/versions=10/valueSize=8/numRangeKeys=1-24       33.5µs ± 2%    34.0µs ± 2%  +1.38%  (p=0.032 n=5+5)
MVCCGet_Pebble/batch=false/versions=10/valueSize=8/numRangeKeys=100-24      115µs ± 1%     115µs ± 1%    ~     (p=0.690 n=5+5)
MVCCGet_Pebble/batch=false/versions=100/valueSize=8/numRangeKeys=0-24       107µs ± 2%     107µs ± 1%    ~     (p=0.421 n=5+5)
MVCCGet_Pebble/batch=false/versions=100/valueSize=8/numRangeKeys=1-24       117µs ± 1%     118µs ± 1%    ~     (p=0.222 n=5+5)
MVCCGet_Pebble/batch=false/versions=100/valueSize=8/numRangeKeys=100-24     209µs ± 4%     208µs ± 1%    ~     (p=1.000 n=5+5)
MVCCGet_Pebble/batch=true/versions=1/valueSize=8/numRangeKeys=0-24         3.82µs ± 1%    3.81µs ± 1%    ~     (p=0.571 n=5+5)
MVCCGet_Pebble/batch=true/versions=1/valueSize=8/numRangeKeys=1-24         6.19µs ± 1%    6.14µs ± 1%    ~     (p=0.310 n=5+5)
MVCCGet_Pebble/batch=true/versions=1/valueSize=8/numRangeKeys=100-24       69.2µs ± 0%    69.6µs ± 1%  +0.70%  (p=0.032 n=5+5)
MVCCGet_Pebble/batch=true/versions=10/valueSize=8/numRangeKeys=0-24        22.1µs ± 1%    22.0µs ± 2%    ~     (p=0.310 n=5+5)
MVCCGet_Pebble/batch=true/versions=10/valueSize=8/numRangeKeys=1-24        27.6µs ± 1%    27.8µs ± 1%    ~     (p=0.222 n=5+5)
MVCCGet_Pebble/batch=true/versions=10/valueSize=8/numRangeKeys=100-24      94.2µs ± 1%    95.7µs ± 1%  +1.57%  (p=0.032 n=5+5)
MVCCGet_Pebble/batch=true/versions=100/valueSize=8/numRangeKeys=0-24        105µs ± 1%     105µs ± 1%    ~     (p=0.841 n=5+5)
MVCCGet_Pebble/batch=true/versions=100/valueSize=8/numRangeKeys=1-24        114µs ± 1%     114µs ± 2%    ~     (p=0.841 n=5+5)
MVCCGet_Pebble/batch=true/versions=100/valueSize=8/numRangeKeys=100-24      188µs ± 3%     190µs ± 1%    ~     (p=0.310 n=5+5)

Reviewable status: complete! 0 of 0 LGTMs obtained (waiting on @erikgrinaker and @RaduBerinde)

[jbowens]

TFTR!