oxidecomputer / oxidecomputer/maghemite
BGP unnumbered doesn't handle active interfaces state changes
Nobody has claimed this yet.
- Dominant language
- Rust
- Stars
- 94
- Forks
- 6
- Avg merge
- 1d 8h
- Merged PRs (30d)
- 12
Description
The BGP unnumbered codebase has some interface lifecycle handling, which was introduced primarily to handle the scenario where nexus sends mgd config for an unnumbered interface which hasn't yet been created in the underlying OS.
However, this implementation has gaps:
- If the OS interface were to be deleted, the active interface doesn't go back into the pending state
- If the OS interface were to have its ifindex / scope_id change (e.g. upon OS interface deletion/re-creation), this is not detected and the cache is not updated (meaning BGP may keep trying to bind its TCP sockets to an old/wrong/invalid/missing ifindex).
These issues haven't been observed/reported yet, but they are real gaps in the codebase which should be handled.
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by tracing the BGP unnumbered interface lifecycle handling from the code that tracks active and pending interfaces, then follow how the interface cache is used for TCP socket binding. Reproduce or test OS interface deletion and deletion/re-creation with a changed ifindex or scope_id; done means deleted interfaces return to pending and changed identifiers update the cache before binding.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- rust
- Domain
- networking
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 48/100