[BACKPORT] Gemini backports to 2025.1 #12043
Open
+48
−72
Conversation
This file contains hidden or 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
- Enable Statement Logger on all workloads statements are logged now in Oracle Cluster, this removes the problem, performance bottlenck of logging to file and gives us better tracability - Removing --fail-fast flag, this flag still exists in gemini but is not used for anything, every workload should migrate to --max-errors-to-store flags, which gives better control if and how gemini will fail Signed-off-by: Dusan Malusev <[email protected]>
Since the SCT sets the default `gemini_seed` to zero in default configration, the check for `if seed is None:` is always false. And zero seed is always applied. Removing `gemini_seed: 0` from default yaml configuration, makes it less confution and allows to deliberately test gemini stability. Supplying the seed is still allows in Jenkins/Argus in the parameters for full repoducibility of the statements and schema Signed-off-by: Dusan Malusev <[email protected]>
Signed-off-by: Dusan Malusev <[email protected]>
In 2.0 gemini, more thing were added to `gemini_cql_requests` metrics, mainly the `method` which contained what statement type is executed. This broke the upgrade tests, as in upgrade test the first value returned from prometheus is used to check `> 0`, but the prometheus ordered the metric labels in alphabetical order, making the metric currently checked to be *almost* always zero, thus failing the check and killing gemini thinking it's not running (but it actually is). Now fetching the `InsertStatement` label makes sure that this check will succeed (maybe not on the first try as it is not yet inserting, but after a couple of try will succeed). Signed-off-by: Dusan Malusev <[email protected]>
CodeLieutenant
added
backport/none
I dont see any discrepancies Why should we even bring #11540 cause it's just a downgrade, and we are moving forward, this one will never introduce any problems inside the source tree |
Because base class (defaults) were not in the context, `filter_subclasses` method was not considering them when evaluating. Fixed by adding base class attrs to `evaluator.context` This is alternative approach to scylladb#11186 which could not be backported due to missing other parts.
|
@CodeLieutenant I've implemented alternative fix (added commit) to nemesis list resolution. Retrying job. |
|
@CodeLieutenant can you tell what's wrong with upgrade test? it was 'yellow' |
CodeLieutenant
removed
the
test-integration
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
As we are considering gemini stable with 2.x release, it's time to backport it to releases.
Testing
PR pre-checks (self review)
backportlabelsReminders
sdcm/sct_config.py)unit-test/folder)These backports are cherry-picked from master and they have to be done manually