Graceful exit on smf context init failure

[protohuf]

During testing I encountered a situation where DNS lookup of the pfcp failed. This caused SMF to attempt to access a non-existent object which crashed the service (SEGFAULT).

I also noticed that InitSmfContext does return nil under some other circumstances but the return value is not checked (which would cause the service to attempt to access an uninitialized smfCtx).

So I added a check to m make sure the context is valid after init and also made DNS errors cause an invalid context state

[Copilot]

Pull Request Overview

This PR improves the SMF service's resilience during initialization by ensuring that an unsuccessful SMF context initialization is detected early, preventing segfaults. Key changes include:

• Adding a nil-check after SMF context initialization in service/init.go to exit gracefully if initialization fails.
• Updating Docker and related dependencies in go.mod.
• Updating the PFCP address resolution error handling in context/context.go to return nil on failure.

Reviewed Changes

Copilot reviewed 3 out of 4 changed files in this pull request and generated 1 comment.

File	Description
service/init.go	Adds a nil-check for the SMF context and logs a fatal error if initialization fails.
go.mod	Updates Docker dependency version and removes deprecated dependencies.
context/context.go	Returns nil when PFCP address resolution fails, marking invalid context initialization.

Comments suppressed due to low confidence (1)

go.mod:41

• [nitpick] It may be beneficial to add a comment or update the commit message to document why the Docker dependency was updated to v25.0.6. This context can help future maintainers understand the rationale behind the dependency changes.

	github.com/docker/docker v25.0.6+incompatible // indirect

[gab-arrobo]

+1

[gab-arrobo]

@thakurajayL, @ghislainbourgeois, what are your thoughts?

[gab-arrobo]

Hi @protohuf,
Quick question... How can we replicate this issue you are specifically encountered? I am asking this because I was able to replicate the DNS lookup error by setting the pfcp.addr to some non-existing name/address. That is, how are you deploying the smf?

net.DNSError {UnwrapErr: error nil, Err: "server misbehaving", Name: "smf"...}

[protohuf]

Hi @gab-arrobo The short version of the story is that we (our organisation) is experimenting and examining the various SD-Core components in order to get a deeper understanding of the project. This particular issue arose as we where individually setting up each component without any containerization (just running the services directly on bare metal).

So there are no direct deployment method that's currently causing this to happen. I just think it's better if a process dies gracefully rather than segfaulting :)

[ghislainbourgeois]

+1

I think with the log level change suggested by @gab-arrobo this should be good to merge.

[gab-arrobo]

+1