Skip to content

v1.3.4

Choose a tag to compare

View all tags
@bekkaze bekkaze released this 04 Nov 23:34
· 5 commits to main since this release
v1.3.4
ebd346e

WHOIS got a thorough going-over after a bug report, and there are install templates for Unraid, Portainer and CasaOS.

Fixes

  • WHOIS failed where the whois command worked. Lookups only ever used IPv4, so a registry that blocks your server's IPv4 made the check time out even though the same query was answered over IPv6. WHOIS now tries both, in the order your system prefers, the way the whois command does. One time budget is shared between them, so a blocked address cannot use up the time the other one needs (#36).

    Docker gives containers only IPv4, whatever the host has. To use this, uncomment the enable_ipv6 block at the end of docker-compose.yml. See WHOIS that times out.

  • .nl and .it records came back nearly empty. Both write a record as indented blocks rather than key: value lines, so the registrar read as empty and the name servers and abuse contact were dropped. .it also uses shorter field names and headings with no colon. Fields inside a contact block now stay with that contact, so the date under "Admin Contact" is no longer taken as the domain's creation date (#36).

  • A refused query looked like an empty record. Registries answer "rate limit exceeded" and the like as plain text on port 43, with nothing marking it an error, and SIDN does so after as little as one query. Those are reported as failed checks now, as is a WHOIS check on a host like mail1.example.nl, which a registry cannot answer at all.

  • One refusal no longer wipes out a good record. When a registry refuses, the last good answer is served for up to four times WHOIS_CACHE_MINUTES, marked as stale with the reason and its age, instead of the asset losing its registrar and expiry date until the registry is willing again.

  • A failed WHOIS check said nothing. Blacklist, SSL and server checks all report their failures on the dashboard and the asset page. WHOIS was the one that did not.

  • Assets could show "Scheduled check overdue" forever. Releases before 1.3.2 accepted any per-asset interval, including negative ones. The scheduler ignores those and checks on time, but the health summary took them at face value.

  • Saving the scheduler settings during a check cycle threw that cycle's results away. The checks had already run; only saving them was skipped.

New

  • Install templates for Unraid, Portainer and CasaOS, in templates/. The Portainer and CasaOS files are self-contained, so they install without a clone of the repository (#31, #32).
  • WHOIS_TIMEOUT_SECONDS (default 30, shared across both address families) and WHOIS_CACHE_MINUTES (default 360). The timeout was fixed at 10 seconds before, and answers were never reused.

Security

  • Vite updated to 6.4.5 (GHSA-8g5g-mxqh-grgj). A development dependency: the frontend image serves the built files with nginx and never runs Vite's dev server, so installs were not affected.

Upgrading: docker compose pull && docker compose up -d. No other steps. Pull the repository too if you want the enable_ipv6 option or the install templates.

Full Changelog: v1.3.3...v1.3.4

Assets 2
Loading