chore: Add FakeRedis capture/compare to more cluster tests and seed during migration #4542
Conversation
Note that we can't add such a test to `test_cluster_native_client` because we don't do a migration there, just re-assign slots, meaning that data is lost (by desgin?) Fixes #4429
This is just a "free" step, which we compare ourselves to FakeRedis. Sort of sanity check to see that we behave like Redis does. We're supposed to take this further and seed during migration, then compare to FakeRedis. So far we didn't do it because we didn't have anything to compare against, as traffic sent during migration was only available to the target node once the migration finished. |
chakaz
changed the title
|
@kostasrim I now added the remainder of the tests (originally I thought about doing so in a separate PR, but I gave up on that due to time considerations). |
tests/dragonfly/cluster_test.py
Outdated
|
|
||
| await nodes[1].client.execute_command( | ||
| "DFLY", | ||
| "LOAD", | ||
| "snap_during_migration-summary.dfs", | ||
| ) | ||
|
|
||
| assert await seeder.compare(capture_before_migration, nodes[1].instance.port) | ||
| # TBD: We can't compare the post-loaded data with anything because it is saved while we seed |
There was a problem hiding this comment.
I think you can still do this check if save will run before you start seeding traffic that will run during the migration.
you will need to first run
await seeder.run(target_deviation=0.1)
to fill the server with data
and than create a capture and compare it at the end
Note that we can't add such a test to
test_cluster_native_clientbecause we don't do a migration there, just re-assign slots, meaning that data is lost (by desgin?)Fixes #4429