test: disable NAT port mapping, outbound dials, inbound connections

[Stebalien]

Proposed Changes

My poor network. It deserves to be treated with respect and dignity and doesn't deserve to be spammed.

1. Disable NAT port mapping. Because no, we don't want our integration tests nodes to be reachable.
2. Disable all but a single localhost/quic transport. No need to do more work than necessary.
3. Set the connection manager limits to be really high. This probably doesn't matter, but there's no need to be killing connections in our integration tests.
4. Reject all outbound dials to non-localhost addresses.

Checklist

Before you mark the PR ready for review, please make sure that:

• Commits have a clear commit message.
• PR title conforms with contribution conventions
• Update CHANGELOG.md or signal that this change does not need it per contribution conventions
• New features have usage guidelines and / or documentation updates in
  • Lotus Documentation
  • Discussion Tutorials
• Tests exist for new functionality or change in behavior
• CI is green

[ribasushi]

this is great - want.

[rvagg]

The ddo test being flaky here is weird (rerun made it pass) and makes me sus. But is that because of #12592?

[Stebalien]

The ddo test being flaky here is weird (rerun made it pass) and makes me sus. But is that because of #12592?

Shouldn't be, this branch is up to date. But I'm also not sure how this change could affect DDO...

[rvagg]

Yeah, just a quick look at the logs it seems like it's something about the miner<>node cooperation in that test. It might not be odd for a flaky, but I don't recall seeing that particular test as flaky, ever.

[Stebalien]

Yeah, just a quick look at the logs it seems like it's something about the miner<>node cooperation in that test. It might not be odd for a flaky, but I don't recall seeing that particular test as flaky, ever.

Yeah, I haven't seen that before. I'll rerun it a few times and see.

[Stebalien]

So, 1 failure for DDO against 4 successes. Looking at the failure, it looks like it took slightly too long:

2024-10-12T14:32:43.3573014Z         	Error:      	Should not be: 199
2024-10-12T14:32:43.3573236Z         	Test:       	TestOnboardRawPieceSnap
2024-10-12T14:32:43.3573713Z         	Messages:   	block at height 4512 never managed to sync to node, which is at height 4511

[Stebalien]

Hm. Although TestOnboardRawPieceSnap is timing out after 10m. Waiting on

lotus/itests/direct_data_onboard_test.go

Line 388 in f6d4fc7

miner.WaitSectorsProving(ctx, toCheck)

.