Skip to content

First-install checklist, just status, robust relay credential permissions - #41

Open
byjpr wants to merge 1 commit into
claude/admiring-planck-25cvzffrom
claude/fix-17-first-install
Open

byjpr wants to merge 1 commit into
claude/admiring-planck-25cvzffrom
claude/fix-17-first-install

Conversation

@byjpr

@byjpr byjpr commented Oct 9, 2026

Copy link
Copy Markdown
Owner

Closes #17

Changes

  • Relay credentials (modules/relay.nix): the rendered postfix-relay-credentials map is now root:postfix 0440 instead of root 0400. It stays unreadable to everyone else, and relaying works whether Postfix's smtp client opens the texthash: table before or after dropping privileges.
  • docs/first-install.md: a verification checklist for a new provider or region:
    • services, networking, decrypted secrets and certificate (including how to retry ACME);
    • just check, test mail in both directions with SPF/DKIM/DMARC in Gmail's "Show original", mail-tester and internet.nl;
    • relay log lines and permissions;
    • the OVH reinstall step;
    • recovering an unreachable server through the provider's rescue system, since there is no root password for the web console.
  • just status (scripts/status.sh): service states, failed units, queue size, certificate expiry, disk, pending reboot and uptime.
  • The README points to the checklist after installation.

This PR can't remove the need to try the remaining untested paths on real servers (OVH reinstall, the first boot per provider). The checklist makes that a defined, quick step.

Testing

  • The sops-nix manifest in the built system shows the template as {"owner":"root","group":"postfix","mode":"0440"}.
  • The ACME unit name in the docs (acme-order-renew-<host>.service) matches the built system.
  • All 5 provider example systems build. nixfmt, shellcheck and tofu validate pass.

🤖 Generated with Claude Code

https://claude.ai/code/session_01G22o7QQ17GRtkCqpJNFr9B


Generated by Claude Code

…ions

- The relay credentials map is now root:postfix 0440 (was root 0400), so
  relaying works however Postfix opens the texthash table.
- docs/first-install.md: what to verify on a new provider or region, how
  to read the relay logs, and how to recover an unreachable server via the
  provider's rescue system.
- just status: services, failed units, queue, certificate expiry, disk and
  pending reboot at a glance.

Closes #17

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01G22o7QQ17GRtkCqpJNFr9B
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.

[Medium] Untested code paths: relay credentials file permissions, OVH reinstall, first boot on each provider

2 participants