[Merged by Bors] - sql: add Database.WithConnection

[ivan4th]

Motivation

In some situations (most notably, syncv2) a lot of database queries
need to be made with no need for a read or write transaction.
Database connection pool adds noticeable delays to the queries.
Using read transactions instead of simple connections has the side
effect of blocking WAL checkpoints.

Description

This adds a possibility to take a connection from the pool to use it
via the Executor interface, and return it later when it's no longer
needed. This avoids connection pool overhead in cases when a lot of
quries need to be made, but the use of read transactions is not
needed.

Additionally, this PR removes unused Rollback method from migrations.
The failed migrations are reverted either by rolling back the transactions
or dropping the temporary database used when vacuuming is needed.

UPD: Rollback method can be useful after all, so reverted this part of the PR.

[codecov]

Codecov Report

Attention: Patch coverage is 54.83871% with 14 lines in your changes missing coverage. Please review.

Project coverage is 79.9%. Comparing base (f711c61) to head (571a53e).
Report is 15 commits behind head on develop.

Files with missing lines	Patch %	Lines
sql/database.go	44.0%	10 Missing and 4 partials ⚠️

Additional details and impacted files

@@            Coverage Diff            @@
##           develop   #6445     +/-   ##
=========================================
- Coverage     79.9%   79.9%   -0.1%     
=========================================
  Files          352     352             
  Lines        46096   46126     +30     
=========================================
+ Hits         36865   36870      +5     
- Misses        7142    7161     +19     
- Partials      2089    2095      +6     

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

---

🚨 Try these New Features:

• Flaky Tests Detection - Detect and resolve failed and flaky tests

[ivan4th]

bors merge

[spacemesh-bors]

Build failed:

• systest-status

[ivan4th]

systest failure:

=== FAIL: systest TestPartition_50_50 (133.53s)
    logger.go:146: 2024-11-13T21:32:41.022Z	INFO	TestPartition_50_50	testcontext/context.go:391	using	***"namespace": "test-iyhh"***
    logger.go:146: 2024-11-13T21:32:41.036Z	INFO	TestPartition_50_50	cluster/cluster.go:166	Using the following nodes	***"total": 20, "bootnodes": 2, "smeshers": 18, "old smeshers": 0, "remote": 0***
    partition_test.go:199: 
        	Error Trace:	/src/systest/tests/partition_test.go:199
        	Error:      	Received unexpected error:
        	            	watcher is terminated while waiting for pod with id smesher-0
        	Test:       	TestPartition_50_50

This looks like stuck network or smth like that.
Given that the new temporary connection mechanism isn't actually used anywhere yet in this PR, this is unrelated.

[fasmat]

I'm not exactly sure what went wrong, but all smeshers were deployed quite late: https://grafana.spacemesh.dev/goto/wJ_klsGHg?orgId=1

smesher-0 came online ~22:34:40 and 1 second later the test runner decided to not try to ping it again and end the test. Normally the test should wait 5 (bootstrap-duration) minutes before giving up, I'm not sure why it either didn't came online within that time or why the runner gave up earlier 🤷

[fasmat]

bors merge

[spacemesh-bors]

Build failed:

• systest-status

[ivan4th]

bors merge

[spacemesh-bors]

Pull request successfully merged into develop.

Build succeeded:

• ci-status
• systest-status