fast import: restore to neondb (not postgres) database

[NanoBjorn]

Problem

postgres is system database at neon, so we need to do pg_restore into neondb instead

https://github.com/neondatabase/cloud/issues/22100

Summary of changes

Changed fast_import a little bit:

1. After succesfull connection creating neondb in postgres instance
2. Changed restore connstring to use new db
3. Added optional source_connection_string, which allows to skip s3_prefix and just connect directly.
4. Added -i that stops process until sigterm

TODO

• test image in cplane e2e
• Change import job image back to latest after this merged (partial revert of https://github.com/neondatabase/cloud/pull/22338)

[github-actions]

7315 tests run: 6939 passed, 0 failed, 376 skipped (full report)

---

Flaky tests (1)

Postgres 17

• test_physical_replication_config_mismatch_max_locks_per_transaction: release-arm64

Code coverage* (full report)

• functions: 33.7% (8420 of 24980 functions)
• lines: 49.2% (70397 of 143103 lines)

* collected from Rust tests only

---

_{The comment gets automatically updated with the latest test results
ebc4735 at 2025-01-15T20:28:25.536Z :recycle:}

[problame]

In addition to the feedback I gave on Slack, I think instead of building a complicated CLI, let's allow specifying the spec as a filesystem path in addition to an S3 location.

It'd be a good Rust learning exercise to do that refactoring, but, probably better to wait for someone from Storage to pick this up / pair with you.

The fast_import.rs code isn't up to our usual standards, remember, it's PoC code.

[NanoBjorn]

I think instead of building a complicated CLI, let's allow specifying the spec as a filesystem path in addition to an S3 location.

Though about it as well, but not sure I like the idea of putting some testing args into the spec. Optional args have the same semantics as top-level json spec values, so I would leave them as is for now (until I really need any of those in the spec)

[hlinnaka]

Hmm, do we want to always restore to 'neondb'? I'd expect the restore to preserve the original database name, i.e. if the source database is called "myapp_staging", I'd expect it to be imported as "myapp_staging" too.