Use random port number when starting cardano-testnet

[carbolymer]

Description

This PR makes cardano-testnet use random port range when starting the testnet. The issue arises in Hydra, when cardano-testnet is executed for multiple architectures at once, using separate invocations of cabal test. This can result in clashes in occupied ports by different testnets and test failure similar to this (taken from https://ci.iog.io/build/3734298/nixlog/1):

Cardano node process did not start: 
cardano-node: DiffusionError Network.Socket.bind: resource busy (Address already in use)
Cardano node process did not start: 
cardano-node: DiffusionError Network.Socket.bind: resource busy (Address already in use)
Cardano node process did not start: 
cardano-node: DiffusionError Network.Socket.bind: resource busy (Address already in use)

Checklist

• Commit sequence broadly makes sense and commits have useful messages
• New tests are added if needed and existing tests are updated. These may include:
  • golden tests
  • property tests
  • roundtrip tests
  • integration tests
    See Runnings tests for more details
• Any changes are noted in the CHANGELOG.md for affected package
• The version bounds in .cabal files are updated
• CI passes. See note on CI. The following CI checks are required:
  • Code is linted with hlint. See .github/workflows/check-hlint.yml to get the hlint version
  • Code is formatted with stylish-haskell. See .github/workflows/stylish-haskell.yml to get the stylish-haskell version
  • Code builds on Linux, MacOS and Windows for ghc-8.10.7 and ghc-9.2.7
• Self-reviewed the diff

Note on CI

If your PR is from a fork, the necessary CI jobs won't trigger automatically for security reasons.
You will need to get someone with write privileges. Please contact IOG node developers to do this
for you.

[smelc]

@carbolymer

when cardano-testnet is executed for multiple architectures at once

Multiple architectures are trying to use the same port number of one machine? How can this happen? Are the architectures being emulated and so are sharing the same pool of ports of one physical machine?

when cardano-testnet is executed for multiple architectures at once, using separate invocations of cabal test. This can result in clashes in occupied ports by different testnets and test failure similar to this

So wouldn't it be simpler to multiply numberOfPorts by the number of architectures?

[carbolymer]

@smelc

Multiple architectures are trying to use the same port number of one machine? How can this happen? Are the architectures being emulated and so are sharing the same pool of ports of one physical machine?

Hydra. It's cross-compiling and running tests on the same host.

So wouldn't it be simpler to multiply numberOfPorts by the number of architectures?

Do you mean to use that instead of maxPortsPerRequest? I'd have to find out and hash the architecture currently used to get different starting port offset too. But it wouldn't protect you against running multiple test suites in different hydra jobs (different cabal test invocations) at the same time.

All in all it's just a hack. You'd have to implement some fancy synchronisation in the operating system state (using files in /run maybe) and coordinate between different cabal test executions. I don't know how much Hydra isolates evaluations and if that's even possible.