[earakely-scale]

Part of FD-5420: a computer's own clock follows the gateway's universe clock. #614 made clock writes able to wait for the computers following the clock (wait_followers_s). This makes that wait stop for a computer that has lost its connection.

The rest of what was #634 (clock writes behind the control token, watches bound to each computer's read token) is deferred with FD-5472, along with moving tokens out of environment variables.

What changes

• A computer that has stopped watching isn't waited for. A follower holds a watch up to 25 s and reconnects at once. One whose last watch is older than that plus 5 s (ClockFollowers.QUIET_S) has lost its connection. A write given wait_followers_s now names such a computer at once instead of waiting out its full time.
• A computer that goes quiet during the wait ends the wait at its quiet deadline: stale_after wakes at each acknowledgement and at the next quiet deadline of a computer it waits for.
• A computer that hasn't watched yet is still waited for, since it may be starting.
• The 504 carries followers: each computer's last shown version and when it last watched.

Without this, once core passes wait_followers_s (FD-5473), a computer whose follower died would hold every clock change on it for the whole wait, up to 30 s.

Testing

• Unit: tst/unit/env/gateway/clock_watch_test.py covers:

  • a computer that stopped watching named at once, with when it last watched, while the other is waited for as before;
  • one that goes quiet during the wait named at its quiet deadline, not after the whole wait;
  • one that hasn't watched yet waited for its full time;
  • the 504's followers.

  All 988 gateway unit tests pass.

• Integration, on an EC2 host (x86_64): gateway_test.py's clock tests pass. These are the clock extension, its rate accuracy, the time triggers, the autonomous trigger and sync_env_clock.

• E2E, on the same host, against the branch's gateway image with two real follower processes long-polling under their read tokens:

  Case	Result
  Arming the clock (no token: the writes stay open)	200
  An advance waiting for both computers	200, in 4 ms
  One computer paused (SIGSTOP) for 36 s, then an advance waiting up to 10 s	504 in 5 ms, naming it, last watched 36 s before; the other acknowledged
  The paused one resumed, then an advance waiting	200, in 4 ms
  The gateway restarted, then a re-arm waiting	200

🤖 Generated with Claude Code

Confidence Score: 5/5

The PR appears safe to merge; no blocking issues were found.

Summary

Clock writes now stop waiting for computers whose last watch is at least 30 seconds old.

• Clock writes stop waiting for computers that have gone quiet.

No actionable issues were found. Tests were inspected, not run.

Diagram

%%{init: {'theme': 'neutral'}}%%
flowchart TD
  W[Clock write] --> C[Wake held watches]
  C --> S{Any followers behind?}
  S -->|No| OK[Return 200]
  S -->|Yes| A{Any still within quiet limit or never watched?}
  A -->|No| E[Return 504 with follower details]
  A -->|Yes| D{Wait deadline reached?}
  D -->|Yes| E
  D -->|No| P[Wait for acknowledgement or next deadline]
  P --> S

Loading

_{Reviews (3) · Last reviewed commit: "feat(gateway): a waiting clock write nam..." · Reviewed by Greptile}