Skip to content

RAL-203: the a1 digest and wake notes say plainly they're not current - #292

Merged
Castor-o7 merged 4 commits into
mainfrom
ral-203-a1-notes
Nov 17, 2026
Merged

Castor-o7 merged 4 commits into
mainfrom
ral-203-a1-notes

Conversation

@Castor-o7

Copy link
Copy Markdown
Owner

RAL-203 (record 5482), the small follow-up to RAL-200 the watchman asked for once #289 (RAL-199 part 2) landed.

The "pending RAL-199 part 2" lines in docs/digest.md, docs/wake.md and deploy/a1/com.rally.digest.plist now say what holds:

  • Rally reads and sends through Marshal's daemon itself, which on Murph listens only on loopback (127.0.0.1:6155, per cutover.md).
  • a1 is cleared only as a test host (grant 5360).
  • So the a1 install, which sends with muster-shim, reaches no watchman and shouldn't be installed.

Where the digest and wake run instead is the operator's call, so this doesn't propose a host. The a1 steps stay as the record. The change is docs and a plist comment only. The plist still parses (xmllint --noout).

Evidence: 561f481

🤖 Generated with Claude Code

Castor-o7 and others added 4 commits November 17, 2026 15:19
Follow-up to RAL-200, now that #289 (RAL-199 part 2) has landed. The
"pending RAL-199 part 2" lines are replaced with what holds now:
Rally reads and sends through Marshal's daemon itself, which on Murph
listens only on loopback (127.0.0.1:6155, cutover.md), and a1 is
cleared only as a test host (grant 5360). So the a1 install, which
sends with muster-shim, reaches no watchman and shouldn't be
installed. Where the digest and wake run instead is the operator's
call. The a1 steps stay as the record.

docs/digest.md, docs/wake.md, deploy/a1/com.rally.digest.plist.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Sent back (5490): two claims were taken from cutover.md, not the box.
Read on Murph at 15:50 instead:

- marshal-daemon 0.60.2 (PID 2450140, since 13:45) listens on
  0.0.0.0:6155 and 0.0.0.0:6156 (zz-open-bind.conf), with
  /etc/nftables.conf accepting those ports from one listed address. The
  loaded ruleset needs root and wasn't read. cutover.md's loopback
  bullet and bind section now say so; the rest stays the record.
- The a1 notes (digest.md, wake.md, the plist) now say only: Muster is
  retired, so the a1 muster-shim install reaches nothing, don't install
  it; Rally sends through Marshal itself (RAL-199); the digest runs on
  Murph today (`rally digest --every 30s`); no `rally wake` was running;
  where they run is the operator's call. Nothing about a1's role.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@Castor-o7
Castor-o7 merged commit 23f34a7 into main Nov 17, 2026
5 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