Skip to content

go/worker/storage/committee: Delay fetching chunk when no peers available #6492

@martintomazic

Description

@martintomazic

Hot loop caused the worker's checkpoint sync to become stuck:

{"caller":"checkpoint_sync.go:108","chunk":161,"err":"no peers given to service the request","level":"error","module":"worker/storage/committee","msg":"failed to fetch chunk from peers","runtime_id":"000000000000000000000000000000000000000000000000f80306c9858e7279","ts":"2026-04-06T15:12:28.315550177Z"}
....  repeat N times

Solution:

  • Add backoff/timeout or per-chunk backoff as suggested.
    • Lets keep it simple for the start and only have simple delay.
  • Fix defer in the for loop (resource leak).

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions