Matt Lord

@brandt There are others better suited for the review, but I just wanted to say THANK YOU for the contribution! It's really nice to see non-PlanetScale contributors to the operator....

@icereed, let me know if you'd like any help. I've been thinking about working on this at some point too. :-)

I don't notice anything obvious in the log. What did the [performance_schema.group*](https://dev.mysql.com/doc/refman/8.0/en/group-replication-monitoring.html) tables show? This was my hobby project in the past and I'm not really working on it anymore...

I've confirmed that I no longer see the traffic pause(s) in the test when adding this to the `vtgate` config: ``` diff --git a/vtgate.yaml b/vtgate.yaml index 636a634..0312aad 100644 --- a/vtgate.yaml...

I could not repeat the issue after many rounds of scaling the consul-server StatefulSet from 0-3,3-0 pods when using this image for the vtgate's that was built w/o the cache...

> The current implementation now reuses same keyspace, sets tablet type to `DRAINED` and ensures not to start eplication. Re-using the same keyspace seems more logical to me at first...

Thank you for the bug report, @L3o-pold ! I noticed that the error was against a replica: ``` SQLSTATE[42S02]: Base table or view not found: 1146 target: commerce.80-.replica: vttablet: rpc...

Thanks, @L3o-pold ! I suspect this may be fixed by the work backported here: https://github.com/vitessio/vitess/pull/10796 That didn't seem to make it into 14.0.2: https://github.com/vitessio/vitess/blob/main/doc/releasenotes/14_0_2_changelog.md Is it reasonable to use the...

I also wanted to mention this great blog post that covers the cutover phase in more detail: https://vitess.io/blog/2022-04-06-online-ddl-vitess-cut-over/ I wonder if the cutover is taking more than 10 seconds OR...

OK, thanks @L3o-pold ! It sounds like you might have a simple test case using the local examples? Could you please share that?