Add: retry failed queries in tck automatically without aborting

[xtcyclist]

Introduction

In some CI jobs, waiting 6 seconds is not long enough for DDLs or queries before them to finish, causing subsequent queries to fail. Common errors reported include something like "tag not found", "vid not found", blablabla.

The root cause is that we do not yet support transactions to wrap up DDLs and act upon them on their successful commits only.

For the time being, we could simply retrying such queries every X seconds for Y times. With more time, they may run successfully and the CI job can go on.

Contents

Related work