Repository navigation
feat(dns): move the apex to the platform and restore external-dns ownership of www - #78
Conversation
Run 43675268687 (source 81ab7a9) created the www ownership record and forgot the www CNAME, moving the state from serial 4 to 5 on the same lineage with the configuration unchanged. The `after` and fresh-runner `recovered` packages authenticate with the independent identity and carry the same state digest. The fresh runner's plan wanted to recreate the ownership record only because external-dns had already taken it over; this pull request forgets it. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
external-dns took over www at 04:06:52Z and wrote its ownership records for the new A and AAAA addresses, but both were gone within a minute. www still points at the gateway, but nothing owns it now. - Create `_externaldns.a-www` and `_externaldns.aaaa-www` with exactly external-dns's content, so it re-owns the existing records without changing them, and forget the one-time cname-www marker. - Add a reviewed handover list, .github/dns-handover.json. A marker may claim a live record nothing here manages only when that list names it; every other claim still has to be a record the same plan forgets. This avoids adopting the www addresses here, which would publish the gateway's origin addresses. Part of #21. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Rollback plan, written before the apply What this apply does: it creates two TXT records, What to check: both records still exist five minutes later. external-dns logs no change to www. www keeps answering with a 301 to the apex. If something goes wrong:
🤖 Generated with Claude Code |
Edit the existing apex CNAME in place from devantler.github.io (DNS only) to www.devantler.tech (proxied). One record changes its target, so the site is never without an address, and the apex then follows www and the gateway. This replaces #77, which carried the same edit. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Rollback plan, written before the apply What this apply does: it edits the apex CNAME in place, from How to check it worked: Rollback, with no gap:
Keep GitHub Pages and its 🤖 Generated with Claude Code |
From the local review of the handover list: - A listed record may only be claimed by its own typed marker (`_externaldns.<type>-<name>`), never by a wider typeless one. - Listed names use hostname characters without `_`, so the list can never name a DKIM or other service record. - Test a listed name with an unlisted type, a typeless marker, an extra top-level key, the 20-entry cap, an underscore name, and the repository's own .github/dns-handover.json; each rule fails a test when removed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
devantler
left a comment
There was a problem hiding this comment.
Local review of #78. A fresh reviewer read the handover list at 511b820, then the head moved to add the apex edit. I reviewed the apex edit and the fixes myself, up to 7cfd080. CodeRabbit is out of its hourly allowance, and Codex and Bugbot are at their usage limits.
What the fresh review confirmed, with no critical or major findings:
- Every malformed list fails closed: missing, empty, not an object, a lowercase type, an extra key, an array where a name belongs.
- Names are normalised the same way on both sides.
- Claims on records declared here are still refused, and so is the typeless apex, even if it is listed. MX and TXT records can never be listed.
apply-dns.shvalidates the file before the check. The intended plan passes on jq 1.8.1.
Three minor findings, fixed in 7cfd080:
- A typeless
_externaldns.wwwcould pass, and it claims more than the list names. A listed record may now be claimed only by its own typed marker. - List names allowed
_, so the list could name a DKIM record. Underscores are no longer allowed. - Tests were missing for a listed name with an unlisted type, an extra top-level key, the 20-entry cap, an underscore name, and the repository's own list file. They are added, and each rule fails a test when removed.
The apex edit is the same in-place change reviewed on #77. The CNAME's content goes from devantler.github.io to www.devantler.tech, and proxying is turned on. The guard admits it, because no external-dns marker owns the apex before or after the change. The rollback plan is in the comments.
scripts/test.sh passes (21 suites), and the DNS tests pass on jq 1.7.1 and 1.8.1. Clean at 7cfd080.
🤖 Generated with Claude Code
|
@coderabbitai review 🤖 Generated with Claude Code |
Action performed
|
|
No actionable comments were generated in the recent review.
|
external-dns's `_externaldns.a-www` and `_externaldns.aaaa-www` records are back in the zone with its own content, so the plan for #78 refused to create or adopt them (external-dns-owned) and applied nothing. Drop the two record resources and empty the handover list; the apex edit and the forget of the old cname-www marker remain. Part of #21. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Why
devantler.tech should be served by the platform, like www now is, and keep following the gateway if its address ever changes. www also needs its external-dns ownership back: right after external-dns took over www, its two ownership records for the new addresses disappeared. www still works, but it would not follow a gateway change.
What
Replaces #77. The rollback plans are in the comments: the apex goes back with an in-place edit to
devantler.github.iowith proxying off, and there is a separate plan for the two ownership records. The platform cutover verifies the apex live and watches the ownership records for five minutes after the apply.Part of #21
🤖 Generated with Claude Code