Container removal broke a lot of tests

[roman-khimov]

Expected Behavior

Like before #2638 merge: https://http.fs.neo.org/HXSaMJXk2g8C14ht8HSi7BBaiYZ1HeWh2xnWPGQCg4H6/1168-1703535611/index.html

Current Behavior

https://http.fs.neo.org/HXSaMJXk2g8C14ht8HSi7BBaiYZ1HeWh2xnWPGQCg4H6/1169-1703535726/index.html

Possible Solution

Fix it.

Steps to Reproduce (for bugs)

Run testcases.

Regression

Yeah, #2638.

[cthulhu-rider]

what prevents us from tracking tests' breakage before merging?

[roman-khimov]

It was ~ok: https://http.fs.neo.org/HXSaMJXk2g8C14ht8HSi7BBaiYZ1HeWh2xnWPGQCg4H6/1165-1703509070/index.html

[cthulhu-rider]

It was ~ok: https://http.fs.neo.org/HXSaMJXk2g8C14ht8HSi7BBaiYZ1HeWh2xnWPGQCg4H6/1165-1703509070/index.html

i see, we run only a bit more than 10% of all tests on PRs. Is there a plan to autorun all tests?

[roman-khimov]

These are just different modes, PR and master. 2h+ test suite for every push into PR is too expensive.

[cthulhu-rider]

can PR tests be run when the changes are ready-to-merge (manually, or on approvals) somehow? all tests take some time anyway, so it's better to catch failures in the PR, not within merge-detect-PR-merge-detect-PR... flow

[roman-khimov]

Only if https://github.com/nspcc-dev/neofs-node/actions/workflows/run-tests.yml is to be parameterized for that (it autodetects PR/master currently). But you'll forget to run it, don't worry.

[cthulhu-rider]

forget to run it, don't worry

forget to run - another problem, dont run at all - current

[cthulhu-rider]

same OKness again https://http.fs.neo.org/HXSaMJXk2g8C14ht8HSi7BBaiYZ1HeWh2xnWPGQCg4H6/1176-1703703416/index.html