Unify rocksdb core thread shutdown and cleanup

[martykulma]

RockDB instance would not clean up when the main loop exited due to shutdown command, but it did in most every other case. For the OOD upsert test, this could result in lgalloc failing to allocate disk space for a file, leading to SIBABRT.

Motivation

fixes https://github.com/MaterializeInc/database-issues/issues/9277

Checklist

• This PR has adequate test coverage / QA involvement has been duly considered. (trigger-ci for additional test/nightly runs)
• This PR has an associated up-to-date design doc, is a design doc (template), or is sufficiently small to not require a design.
• If this PR evolves an existing $T ⇔ Proto$T mapping (possibly in a backwards-incompatible way), then it is tagged with a T-proto label.
• If this PR will require changes to cloud orchestration or tests, there is a companion cloud PR to account for those changes that is tagged with the release-blocker label (example).
• If this PR includes major user-facing behavior changes, I have pinged the relevant PM to schedule a changelog post.

[martykulma]

@def- I ran this test locally a few times, but I'm assuming that it's not readily reproducible. If there's any additional testing that I should run, let me know and I'll try it out.

[def-]

The change itself makes sense to me. We will see in a few days from ci-failures (https://ci-failures.dev.materialize.com/?key=test-failures&tfFilters=%5B%7B%22id%22%3A%22build_date%22%2C%22value%22%3A%5Bnull%2Cnull%5D%7D%2C%7B%22id%22%3A%22test_suite%22%2C%22value%22%3A%22Upsert%22%7D%5D&tfGlobalFilter=sig%3D7) if this SIGBUS is still happening.