The project currently has two disjoint TCL frameworks: - Cluster framework, https://github.com/redis/redis/tree/unstable/tests/cluster/tests, which pre-allocates N nodes and uses them for testing large configurations - Normal testing framework, https://github.com/redis/redis/tree/unstable/tests/unit (also integration and modules), which individually launches nodes for testing.
The normal TCL testing framework is much more readily tested and is also automatically run as part of the CI for new PRs. Some tests have already been migrated, such as in https://github.com/redis/redis/pull/10920. Creating this issue as a backlog in case someone would like to help the project by migrating additional tests.
Comment From: Raman-Kumar
Okay, I am doing this.
Comment From: madolson
@Raman-Kumar Thanks for offering to help. Please send a couple of smaller PRs incrementally moving the tests. I would also say anyone that picks this should focus on the higher tests (14 - 29) as they are more focused on the external API as opposed to in the internal cluster implementation details.
Comment From: enjoy-binbin
https://github.com/enjoy-binbin/redis/commit/0b80c60b289e7b44240397c0c85ccc24353550c6
daily: https://github.com/enjoy-binbin/redis/actions/runs/7483641294
just want to comment first, i have migrated intermittently before, and the current effect is this. There are some parts that I didn't think of a good way to deal with, such as kill server. and the some of the daily can not pass.
@madolson maybe when you have some times, you can review it and give me some tips, or can take it from me and finish it.
ok i see maybe we should still put replica-migration-2 / replica-migration-3 in the old cluster tests. they are slow in valgrind