tabledesc_test: deflake TestIndexDoesNotStorePrimaryKeyColumnMixedVersion #116791
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
This change is |
rafiss
requested changes
Dec 19, 2023
Xiang-Gu
force-pushed
the
deflake/TestIndexDoesNotStorePrimaryKeyColumnMixedVersion
branch
from
499641c to
6db9546
Compare
rafiss
requested changes
Dec 19, 2023
| @@ -115,6 +114,8 @@ func TestIndexDoesNotStorePrimaryKeyColumnMixedVersion(t *testing.T) { | |||
|
|
|||
| // Assert cluster version upgrade is blocked. | |||
| require.Equal(t, [][]string{{"1000023.1"}}, tdb.QueryStr(t, "SHOW CLUSTER SETTING version;")) | |||
| // disable AOST for upgrade precondition check to ensure it sees the injected, corrupt descriptor. | |||
| upgrades.TestingSetFirstUpgradePreconditionAOST(false) | |||
There was a problem hiding this comment.
this should be:
defer upgrades.TestingSetFirstUpgradePreconditionAOST(false)()
…sion Previously, this test would flake under `--stress` and the root cause is because the upgrade precondition check that it relies on actually checks for corrupt descriptor at AOST '-10s', and it's possible that the corrupt descriptor was injected within the last 10s to the system, and thus the check produced a "false negative" (meaning it saw there isn't no corruption at AOST '-10s' and mistakenly assumed there isn't corruption now) and thus mistakenly allowed the cluster upgrade to proceed. The AOST '-10s' trick in the upgrade was implemented to help with performance, and, for now, I think that outweighs the possibility of producing such false-negatives, so this fix instead modified the test such that it waits 10s after injecting the corrupt descriptor to ensure it will be seen by the time the upgrade logic runs. The downside that comes with this decision is this test will now take ~10-11s to run and we need to skip it under `--stress` (bc 1000 iterations / # of parallelism * 10s can still be larger than 1h, the default timeout for nightly stresses). Release note: None
Xiang-Gu
force-pushed
the
deflake/TestIndexDoesNotStorePrimaryKeyColumnMixedVersion
branch
from
6db9546 to
de78ad1
Compare
rafiss
approved these changes
Dec 20, 2023
|
bors r+ |
|
Build succeeded: |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Previously, this test would flake under
--stressand the root cause is because the upgrade precondition check that it relies on actually checks for corrupt descriptor at AOST '-10s', and it's possible that the corrupt descriptor was injected within the last 10s to the system, and thus the check produced a "false negative" (meaning it saw there isn't no corruption at AOST '-10s' and mistakenly assumed there isn't corruption now) and thus mistakenly allowed the cluster upgrade to proceed.Fixes #116622
Release note: None