[rmcdaniel]

The deadline-health cells runner checks that each cell's server and database containers stopped. A container that is already gone counts as stopped. It matched only Docker's No such object, but on the Ubuntu 26 runner image the Docker CLI prints error: no such object: …. A SQLite cell has no database container, so it halted with "cleanup unconfirmed" even though its checks passed.

Seen in run 41338240707 on #925, on runner image ubuntu26/20261025.179. The same job passes on other runner images, so failures depend on which image a run gets.

• The absence match in codec-deadline-cells.sh is now case-insensitive.
• ServerParityCodecDeadlineCellsTest has a new case: the fake Docker answers in lower case, and every cell must run and pass.

Checks run: the test file passes 9/9. With the original script the new case fails and the other 8 pass.

🤖 Generated with Claude Code

[rmcdaniel]

Security bounded CI-safety review: no finding at exact head 4b2eaf0afc9c3cf21ab113b59515bfd70736a7d4; no Security hold.

The production condition is unchanged except for ASCII case-folding of the two already-admitted Docker absence phrases (no such object / no such container). Other docker inspect failures still emit Could not confirm..., keep cell state, stop later cells, and fail the runner. The existing INSPECT_FAILS control continues to prove a daemon-connect failure is not converted into absence, while the added control covers the observed Ubuntu 26 lower-case output and requires every cell to run successfully.

Independent source checks: exact delta inspection, git diff --check, and Bash syntax for the changed runner passed. I did not run repository test tooling on the lab host. Fresh hosted checks remain the owner gate.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud