vtcombo: close query service on drop database

[brendar]

Description

Closes the query service on DROP DATABASE, which will close connection pools.

Related Issue(s)

Fixes #16605

Checklist

• "Backport to:" labels have been added if this change should be back-ported to release branches
• If this change is to be back-ported to previous releases, a justification is included in the PR description
• Tests were added or are not required
• Did the new or modified tests pass consistently locally and on CI?
• Documentation was added or is not required

Deployment Notes

[vitess-bot]

Review Checklist

Hello reviewers! 👋 Please follow this checklist when reviewing this Pull Request.

General

• Ensure that the Pull Request has a descriptive title.
• Ensure there is a link to an issue (except for internal cleanup and flaky test fixes), new features should have an RFC that documents use cases and test cases.

Tests

• Bug fixes should have at least one unit or end-to-end test, enhancement and new features should have a sufficient number of tests.

Documentation

• Apply the release notes (needs details) label if users need to know about this change.
• New features should be documented.
• There should be some code comments as to why things are implemented the way they are.
• There should be a comment at the top of each new or modified test to explain what the test does.

New flags

• Is this flag really necessary?
• Flag names must be clear and intuitive, use dashes (-), and have a clear help text.

If a workflow is added or modified:

• Each item in Jobs should be named in order to mark it as required.
• If the workflow needs to be marked as required, the maintainer team must be notified.

Backward compatibility

• Protobuf changes should be wire-compatible.
• Changes to _vt tables and RPCs need to be backward compatible.
• RPC changes should be compatible with vitess-operator
• If a flag is removed, then it should also be removed from vitess-operator and arewefastyet, if used there.
• vtctl command output order should be stable and awk-able.

[codecov]

Codecov Report

Attention: Patch coverage is 0% with 4 lines in your changes missing coverage. Please review.

Project coverage is 68.98%. Comparing base (cc68dd5) to head (c903014).
Report is 39 commits behind head on main.

Files	Patch %	Lines
go/vt/vtcombo/tablet_map.go	0.00%	4 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #16606      +/-   ##
==========================================
+ Coverage   68.85%   68.98%   +0.12%     
==========================================
  Files        1557     1562       +5     
  Lines      199891   200693     +802     
==========================================
+ Hits       137644   138457     +813     
+ Misses      62247    62236      -11     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[mattlord]

Thanks, @brendar ! This change makes sense to me there:

err := tablet.qsc.QueryService().Close(ctx)

We should, however, add some kind of test that demonstrates the problem on main and in turn verifies that it's been fixed in the PR branch.

[brendar]

@mattlord I've added assertions to the vtcombo recreate test. It fails on main:

    recreate_test.go:119:
        	Error Trace:	/Users/brendan/src/github.com/Shopify/vitess/go/test/endtoend/vtcombo/recreate/recreate_test.go:119
        	Error:      	Max difference between 16 and 27 allowed is 5, but difference was -11
        	Test:       	TestDropAndRecreateWithSameShards
        	Messages:   	not within allowable delta: mysqlConnCountBefore=16, mysqlConnCountAfter=27
--- FAIL: TestDropAndRecreateWithSameShards (1.87s)

When testing the fix locally I sometimes saw a difference of 1 connection before and after, which I believe is due to some short lived or lazily initialized connection, so I gave it what I think is a pretty wide delta (5 connections).

The alternative to counting connections would be to parse the *PoolMaxCap stats from /debug/vars which felt a bit convoluted but might avoid any flakiness.

[mattlord]

IMO this is a better query as the process list will show internal threads too (and it's deprecated 🙂):

mysql> show processlist;
+----+-----------------+-----------------+-------------+------------------+------+-----------------------------------------------------------------+------------------+
| Id | User            | Host            | db          | Command          | Time | State                                                           | Info             |
+----+-----------------+-----------------+-------------+------------------+------+-----------------------------------------------------------------+------------------+
|  5 | event_scheduler | localhost       | NULL        | Daemon           |  467 | Waiting on empty queue                                          | NULL             |
| 13 | vt_dba          | localhost       | NULL        | Sleep            |    1 |                                                                 | NULL             |
| 23 | vt_dba          | localhost       | vt_commerce | Sleep            |  462 |                                                                 | NULL             |
| 24 | vt_app          | localhost       | vt_commerce | Sleep            |  463 |                                                                 | NULL             |
| 26 | vt_allprivs     | localhost       | vt_commerce | Sleep            |  463 |                                                                 | NULL             |
| 27 | vt_repl         | 127.0.0.1:50890 | NULL        | Binlog Dump GTID |  463 | Source has sent all binlog to replica; waiting for more updates | NULL             |
| 28 | vt_repl         | 127.0.0.1:50889 | NULL        | Binlog Dump GTID |  463 | Source has sent all binlog to replica; waiting for more updates | NULL             |
| 32 | vt_app          | localhost       | vt_commerce | Sleep            |   23 |                                                                 | NULL             |
| 39 | vt_app          | localhost       | vt_commerce | Sleep            |  461 |                                                                 | NULL             |
| 42 | root            | localhost       | vt_commerce | Query            |    0 | init                                                            | show processlist |
+----+-----------------+-----------------+-------------+------------------+------+-----------------------------------------------------------------+------------------+
10 rows in set, 1 warning (0.00 sec)

mysql> show warnings;
+---------+------+-------------------------------------------------------------------------------------------------------------------------------------------+
| Level   | Code | Message                                                                                                                                   |
+---------+------+-------------------------------------------------------------------------------------------------------------------------------------------+
| Warning | 1287 | 'INFORMATION_SCHEMA.PROCESSLIST' is deprecated and will be removed in a future release. Please use performance_schema.processlist instead |
+---------+------+-------------------------------------------------------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)

mysql> select variable_value from performance_schema.global_status where variable_name="threads_connected";
+----------------+
| variable_value |
+----------------+
| 9              |
+----------------+
1 row in set (0.00 sec)

[mattlord]

Nit, but you can do all of this together:

return result.Rows[0][0].ToInt()

[mattlord]

👍 I'm assuming that this fails on main? I think that's what your comment was demonstrating but just triple checking.

[brendar]

Yes, it fails on main.

[brendar]

@mattlord Thanks for the thoughtful review. Comments have been addressed and I've confirmed the test still fails on main.

[mattlord]

Thanks, @brendar !