Skip to content

No single point of failure for status, node board or bootstrap - #59

Merged
ralyodio merged 3 commits into
masterfrom
feat/aggregator-resilience
Oct 21, 2026
Merged

ralyodio merged 3 commits into
masterfrom
feat/aggregator-resilience

Conversation

@ralyodio

Copy link
Copy Markdown
Collaborator

Why

Everything on c0mpute.com that shows or finds nodes depended on dev2 alone:

  • one status aggregator inside the web container, never restarted after it exits
  • one seed
  • bootstrap.json served by that same host
  • no compiled-in seeds

What

  • A second seed and status aggregator are already running on vienna (c0mpute-seed.service, peer 12D3KooWN99F7k4Fj6SdUqZ2VG4ZBLuw1e1ftwdPDBx3GsiDXzqv, /ip4/152.53.47.37/tcp/46339, read-only status on :8091).
  • lib/aggregators.ts: /status, /nodes and the home count query every URL in STATUS_AGGREGATOR_URLS and use the one that sees the most workers.
  • The in-container aggregator restarts on exit.
  • bootstrap::DEFAULT_SEEDS compiles both seeds into the worker and the aggregator; /bootstrap.json falls back to them.
  • infernet_systemd_managed() also checks systemctl --user (infernet v0.1.62 installs a user unit).

Companion: InfernetProtocol/infernet-protocol#51 (self-restarting nodes, Infernet aggregators, provider details).

Checks

tsc --noEmit is clean. Rust is checked in CI (this box has no C linker).

🤖 Generated with Claude Code

c0mpute.com/status, /nodes, the home count and new-node bootstrap all hung
off one unsupervised aggregator process in the dev2 web container and one
seed on dev2, with no compiled-in seeds.

- A second seed + status aggregator now runs on vienna (netcup, p2p 46339,
  read-only status on :8091).
- The site reads every aggregator in STATUS_AGGREGATOR_URLS in parallel and
  uses the one that sees the most workers; the in-container aggregator is
  restarted when it exits (backoff 2 s..60 s).
- Both seeds are compiled in (bootstrap::DEFAULT_SEEDS) for the worker and
  the aggregator, and /bootstrap.json serves them when its env var is empty.
- The worker treats a user-level infernet.service as managing the daemon
  (infernet v0.1.62 installs one), so the two never start rival daemons.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@github-actions

Copy link
Copy Markdown

ThreatCrush Security Scan

9 finding(s)

MEDIUM: 9

Severity Rule Location
MEDIUM js-unescaped-html-sink apps/web/src/app/blog/[slug]/page.tsx:53
MEDIUM js-unescaped-html-sink apps/web/src/app/layout.tsx:105
MEDIUM js-unescaped-html-sink apps/web/src/components/pit-footer.tsx:17
MEDIUM sh-remote-script-execution scripts/dev-setup.sh:25
MEDIUM sh-eval-expansion scripts/dev-setup.sh:35
MEDIUM sh-remote-script-execution scripts/install.sh:162
MEDIUM sh-remote-script-execution scripts/install.sh:187
MEDIUM sh-remote-script-execution scripts/install.sh:305
MEDIUM sh-remote-script-execution scripts/install.sh:322

Snippets are redacted; ThreatCrush never prints matched credential material.

ralyodio and others added 2 commits October 21, 2026 14:10
…eckout too

bbs ran a bun daemon from ~/.infernet/source for a week next to
infernet.service; the reaper only matched the npm path.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@ralyodio
ralyodio merged commit 228fcba into master Oct 21, 2026
12 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant