sqlstats: add retries to stats test on locked table #119977
Merged
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
Previously, the `TestSQLStatsReadLimitSizeOnLockedTable` test would fail very occasionally due to a rare scenario. When stats are written, they contain a column that's a hashed shard index. It's expected that statements are uniformly distributed across this shard, but that's not guaranteed. Later in the test we check a random shard to make sure its stats count exceeds a minimum of 1 (because we place a total lower bound of 8 in the cluster setting, which is then divided by 8 to derive the per-shard limit). This case will occasionally fail if the random shard that's picked happens to only contain a single statement within. This change modifies the loop at the end of the test to expect a `false` value and make sure to get at least a single `true` result after 3 iterations, instead of requiring 3 `true` results every single time. The requirement that the queries run despite contention will still stand since we'll return an error in that case. Resolves: cockroachdb#119067 Epic: None Release note: None
|
This change is |
xinhaoz
approved these changes
Mar 6, 2024
| @@ -705,7 +704,7 @@ func TestSQLStatsReadLimitSizeOnLockedTable(t *testing.T) { | |||
|
|
|||
| // Set table size check interval to .0000001 second. So the next check doesn't | |||
| // use the cached value. | |||
| sqlConn.Exec(t, "SET CLUSTER SETTING sql.stats.limit_table_size_check.interval='.0000001s'") | |||
| persistedsqlstats.SQLStatsLimitTableCheckInterval.Override(ctx, &s.ClusterSettings().SV, time.Nanosecond) | |||
There was a problem hiding this comment.
Why did we need to override this instead of using the set statement above?
There was a problem hiding this comment.
The new code does the same thing. Override just sets the setting value directly.
This is an unnecessary change but one that simplifies test execution by removing the need for a full SQL parse/exec loop when we have the ability to set the value directly.
dhartunian
added
the
backport-23.2.x
|
TFTR bors r=xinhaoz |
|
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, the
TestSQLStatsReadLimitSizeOnLockedTabletest would fail very occasionally due to a rare scenario. When stats are written, they contain a column that's a hashed shard index. It's expected that statements are uniformly distributed across this shard, but that's not guaranteed. Later in the test we check a random shard to make sure its stats count exceeds a minimum of 1 (because we place a total lower bound of 8 in the cluster setting, which is then divided by 8 to derive the per-shard limit). This case will occasionally fail if the random shard that's picked happens to only contain a single statement within.This change modifies the loop at the end of the test to expect a
falsevalue and make sure to get at least a singletrueresult after 3 iterations, instead of requiring 3trueresults every single time. The requirement that the queries run despite contention will still stand since we'll return an error in that case.Resolves: #119067
Epic: None
Release note: None