-
Notifications
You must be signed in to change notification settings - Fork 4k
catalog/lease: fix flake for TestTableCreationPushesTxnsInRecentPast #115046
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
catalog/lease: fix flake for TestTableCreationPushesTxnsInRecentPast #115046
Conversation
28a1765 to
3de9757
Compare
rafiss
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Reviewable status:
complete! 0 of 0 LGTMs obtained (waiting on @fqazi)
-- commits line 9 at r1:
just for my own education: could you say more about why this causes the desired behavior?
Previously, TestTableCreationPushesTxnsInRecentPast could flake because we attempted to increase the uncertainty interval error by adding a delay on KV RPC calls to make the uncertainty interval larger. This would work based on the observed timestamp having a variance between nodes. This wasn't effective, and we could still have intermittent failures, because the variance was not large enough. To address this, this patch will increase the maximum offset on the HLC clock itself, which is used as the base value for computing the uncertainty interval (see computeIntervalForTxn). With this change, we expect a more deterministic reproduction of the txn getting pushed because of the uncertainty interval. Fixes: cockroachdb#114366 Release note: None
3de9757 to
1fb0eac
Compare
fqazi
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Reviewable status:
complete! 0 of 0 LGTMs obtained (waiting on @rafiss)
Previously, rafiss (Rafi Shamim) wrote…
just for my own education: could you say more about why this causes the desired behavior?
I updated the commit message for why it should work
rafiss
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
nice thanks for the explanation!
Reviewable status:
complete! 0 of 0 LGTMs obtained
|
@rafiss TFTR! bors r+ |
|
Build succeeded: |
|
blathers backport 23.2 |
|
blathers backport 23.2.0-rc |
Previously, TestTableCreationPushesTxnsInRecentPast could flake because we attempted to increase odds of hitting the uncertainty interval error by adding a delay on KV RPC calls. This wasn't effective and could still intermittent failures, and instead we are going to directly modify the uncertainty interval by setting a large MaxOffset on the clock. This will cause the desired behaviour in a more deterministic way.
Fixes: #114366
Release note: None