mysql80: Mysql2::Client session_track fails on mysql >= 8.0.20 or 8.0.21?

[junaruga]

I like to open the ticket of this test failure on mysql80 case, coming from #1154 (comment) . Actually I see this issue happens on the current Travis too.

Failures:
  1) Mysql2::Client session_track returns multiple session track type values when available
     Failure/Error: expect(res).to eq(["________"])

       expected: ["________"]
            got: nil
     
       (compared using ==)
     # ./spec/mysql2/client_spec.rb:1053:in `block (3 levels) in <top (required)>'

Seeing the history of builds.

• With mysql-server 8.0.19: ok : Travis log.
• With mysql-server 8.0.21: failure : Travis log.
• With mysql-server 8.0.23: failure : Travis log.

[junaruga]

The changes in the mysql 8.0.20 or 8.0.21 might cause this issue.
https://dev.mysql.com/doc/relnotes/mysql/8.0/en/news-8-0-20.html
https://dev.mysql.com/doc/relnotes/mysql/8.0/en/news-8-0-21.html

[junaruga]

You can see the reproducer of this issue on the following cases in GitHub Actions.

• ubuntu-20.04 ruby 2.4 mysql80 (MySQL 8.0.23) case here.
  https://github.com/brianmario/mysql2/runs/1973366419?check_suite_focus=true
• macos ruby 2.4 (latest) mysql (MySQL 8.0.23) case here. The case will be merged on the PR GitHub Actions: Add macOS cases. #1170 .
  https://github.com/brianmario/mysql2/pull/1170/checks?check_run_id=1987005817

[sodabrew]

Specifically this release note looks like the culprit:

The Performance Schema collected session-related statistics for errors that can occur only globally and not per session. This is no longer done, reducing memory overhead for error instrumentation. Additionally, rows for global errors are no longer included in error summaries reported per thread, account, user, or host. (Bug #30311574)

[junaruga]

It seems it was fixed. See #1220 (comment) .

[Faq]

As this is closed, noone updated those comments:

mysql2/.github/workflows/build.yml

Line 14 in 890342a

# Use ubuntu-18.04 instead of ubuntu-20.04 temporarily, due to a failing test on mysql 8.0.