storage: fail commands outside the replica GC threshold

[mjibson]

This change is

[mjibson]

This is blocked on a decision by #6985.

[bdarnell]

#6985 was decided in favor of allowing the inconsistency, on the grounds that it is self-healing. This change could introduce non-self-healing inconsistencies. However, queries near the GC cutoff would be so buggy prior to this change that I'm not sure if we care.

LGTM

Previously, mjibson (Matt Jibson) wrote…

This is blocked on a decision by #6985.

---

Review status: 0 of 2 files reviewed at latest revision, 1 unresolved discussion, all commit checks successful.

---

storage/replica_test.go, line 5640 [r1] (raw file):

  // Verify a later Get works.
  if _, err := tc.SendWrappedWith(roachpb.Header{
      Timestamp: makeTS(3, 0),

Put the timestamp(3,0) in a variable to make sure it's the same both times it's used.

---

Comments from Reviewable

[tbg]

Previously, bdarnell (Ben Darnell) wrote…

#6985 was decided in favor of allowing the inconsistency, on the grounds that it is self-healing. This change could introduce non-self-healing inconsistencies. However, queries near the GC cutoff would be so buggy prior to this change that I'm not sure if we care.

LGTM

---

Review status: 0 of 2 files reviewed at latest revision, 2 unresolved discussions, all commit checks successful.

---

storage/replica_test.go, line 5642 [r1] (raw file):

      Timestamp: makeTS(3, 0),
  }, &gArgs); err != nil {
      t.Fatalf("could not get data: %s", err)

Add similar tests for a ConditionalPut. I expect this to work (since executeBatch is called from executeWriteBatch), but it should be asserted.

---

Comments from Reviewable

[tbg]

Reviewed 2 of 2 files at r1.
Review status: all files reviewed at latest revision, 2 unresolved discussions, all commit checks successful.

---

Comments from Reviewable

[mjibson]

Review status: 1 of 2 files reviewed at latest revision, 2 unresolved discussions, some commit checks pending.

---

storage/replica_test.go, line 5640 [r1] (raw file):

Previously, bdarnell (Ben Darnell) wrote…

Put the timestamp(3,0) in a variable to make sure it's the same both times it's used.

Done.

---

storage/replica_test.go, line 5642 [r1] (raw file):

Previously, tschottdorf (Tobias Schottdorf) wrote…

Add similar tests for a ConditionalPut. I expect this to work (since executeBatch is called from executeWriteBatch), but it should be asserted.

Done.

---

Comments from Reviewable

[tbg]

Previously, tschottdorf (Tobias Schottdorf) wrote…

:lgtm:

---

Reviewed 1 of 1 files at r2.
Review status: all files reviewed at latest revision, 1 unresolved discussion, all commit checks successful.

---

Comments from Reviewable