Failure modes

[Synicix]

Depends on #70

Summary:

• Fix issue with command parsing to match docker cli command parsing
• Add failure handling for launching a pod with a bad command
• Add failure handling for a pod failing after starting (running failure)
• Moved some sections of list container to a helper function

[codecov]

Codecov Report

Attention: Patch coverage is 97.46835% with 4 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/core/orchestrator/docker.rs	97.47%	3 Missing ⚠️
src/core/error.rs	66.66%	1 Missing ⚠️

📢 Thoughts on this report? Let us know!

[Synicix]

Updated with new code on dev. Will probably hold up on merging until uffi python stuff is in, since that changes a few things.

[guzman-raphael]

Depends on #88

[guzman-raphael]

@Synicix Nice work. 👍

I would prefer if we split off the logs feature into its own dedicated PR to make it easier to review and ensure the feature is complete. Also, the logs feature is missing how to get_logs(..) from a PodRun like the issue mentions (it would go in the Orchestrator trait). You are currently only capturing it after the container terminates.

[Synicix]

@guzman-raphael

Per your suggestion, all logging stuff has been moved to #93

Both are ready for review now.

[guzman-raphael]

Thanks @Synicix. 🤝

Quick update: I've skimmed through this PR but have yet to do a deep dive.

From a quick glance, it does look like it might take some time on the back-and-forth to find the best approach to the problems you've tackled here. Meeting together would help but still would require a reasonable time commitment. Given what it adds, I'm treating this as low priority for now purely until we hit the primary goals of our current milestone.

[Synicix]

Done:
Moved to Crypto PR:

• crypto.rs edit is fine pull request (

Logging PR:

• Unknown -> Unset (Decided against it due to the docker api container status return behavior)

Pending:

• Remove result large err (don't return error)

• Remove regex command parsing (Write issue to deal with chaining and more complicated commands. Change command to a vector of String)

• Change prepare container start inputs change pub (crate)

• Remove delete container duplicate

• get_info() assigned name into separate issue (issues)

• Remove print wait container error

• Remove queued container

• Nested dir hash test should be move to a separate pull request

• fail_at_start() add comment

• Write a function to extract the container name from error 198

• 283 rename list_Result to pod_runs

• integrate the two wrappers in tests

In summary it will create 3 pull requests.

[Synicix]

Closing due to splitting of pull request into smaller ones.