Replaces #89/#261; same change, history without a built binary. This PR is one commit on current main. It also ignores cmd/mesh-controller/mesh-controller, the file go build leaves there.
What
A new row S21 in the signals table adds the condition apply-failing on the machine. It is raised once a machine's reports have said the same failure (refused whole, or applied only in part) for longer than one apply interval (5 minutes, the node-engine's ReconcileEvery), over at least two reports. It is not raised while the machine is silent or asleep, because S1 covers that, and a clean apply clears it. It reads the failing-since and count the store already keeps (issue 065, NotDoingWhatTheyWereTold), so there is no schema change.
When the node-engine names a package manager that does not answer (key the package manager (pacman), mesh-host PR on the same branch name), the condition names the tool in the summary and the headline (" cannot install packages"). The probe's words go in the evidence (Said), so the file path stays out of the summary. Otherwise it gives the refusal or the failed resource ids, with the machine's words in the evidence. Plain words are registered as a fallback.
Why
Issue 233, question 3, decided 2026-10-11: a machine whose declarations are refused, or applied only in part, for longer than one apply interval raises a condition by itself (issue 187's family). On 2026-10-04 a workstation refused every declaration for 90 minutes, and the only sign was a line in status.
How it was tested
apply_failing_test.go: a machine whose package manager has not answered for 90 minutes is raised. The test checks the key, that the summary names pacman and the count and carries no path, that the evidence carries the probe's words, and the plain words. A whole refusal is also raised. A single report and a silent machine are not raised.
The generated signals-table test walks S21 inside and past its bound, raised and then cleared through the keeper. The plain-words test holds its wording.
sh merge-check.sh: green on the change (the store tests were skipped locally because no MESH_TEST_POSTGRES is set). Targeted tests re-run on current main.
Fixes issue 233
Replaces #89/#261; same change, history without a built binary. This PR is one commit on current main. It also ignores `cmd/mesh-controller/mesh-controller`, the file `go build` leaves there.
## What
A new row S21 in the signals table adds the condition `apply-failing` on the machine. It is raised once a machine's reports have said the same failure (refused whole, or applied only in part) for longer than one apply interval (5 minutes, the node-engine's ReconcileEvery), over at least two reports. It is not raised while the machine is silent or asleep, because S1 covers that, and a clean apply clears it. It reads the failing-since and count the store already keeps (issue 065, `NotDoingWhatTheyWereTold`), so there is no schema change.
When the node-engine names a package manager that does not answer (key `the package manager (pacman)`, mesh-host PR on the same branch name), the condition names the tool in the summary and the headline ("<machine> cannot install packages"). The probe's words go in the evidence (`Said`), so the file path stays out of the summary. Otherwise it gives the refusal or the failed resource ids, with the machine's words in the evidence. Plain words are registered as a fallback.
## Why
Issue 233, question 3, decided 2026-10-11: a machine whose declarations are refused, or applied only in part, for longer than one apply interval raises a condition by itself (issue 187's family). On 2026-10-04 a workstation refused every declaration for 90 minutes, and the only sign was a line in `status`.
## How it was tested
- `apply_failing_test.go`: a machine whose package manager has not answered for 90 minutes is raised. The test checks the key, that the summary names pacman and the count and carries no path, that the evidence carries the probe's words, and the plain words. A whole refusal is also raised. A single report and a silent machine are not raised.
- The generated signals-table test walks S21 inside and past its bound, raised and then cleared through the keeper. The plain-words test holds its wording.
- `sh merge-check.sh`: green on the change (the store tests were skipped locally because no MESH_TEST_POSTGRES is set). Targeted tests re-run on current main.
Fixes issue 233
mesh/merge-gate pass: builds build-agent, mesh-controller → ace, g14, novox, shanks; no bus step; every machine composes with the change as it did without …
composed together: pass — every machine composes with the change as it did without (4 of 4 compose)
Machines
mesh-controller on novox: passed — healthy 3 times over 4m7s
mesh-controller on the rest running it: sent
Transitions
2026-10-12 13:39 (new) → proposed (announced): novox/mesh-controller#263's head announced
2026-10-12 13:48 proposed → checked (checked): the verdict names this commit
2026-10-12 13:48 checked → ready (accepted): the gate passed or warned, and the repository's own check did not fail
2026-10-12 14:11 ready → published (merged): on the trunk its modules follow: merged there, and its walk opened — or nothing for a walk to move
2026-10-12 14:11 published → delivering (go): its walk started: let go by this owner in its group's order, by a person, or on the controller's own path — not a failed walk that carried a later commit for it, after which it waits for a walk of its own commit
2026-10-12 14:16 delivering → delivered (done): its walk is done, carrying its own commit: every machine of its deploy plan passed, or was left as its policy says
The commit's note under refs/notes/mesh-plan keeps every transition: git log --notes=mesh-plan.
<!-- mesh-delivery:view -->
**Delivery** `novox/mesh-controller@d4de4543ccbf` — **delivered** since 2026-10-12T14:16:30Z
**Delivery plan** — builds mesh-controller → novox; no bus step
- tier 0: mesh-controller
- novox: receives mesh-controller
**Group** `fix/233-a-failed-probe-fails-only-what-needs-the-tool-2`, in order: novox/mesh-host@74ed7cbe0ab4 → novox/mesh-controller@d4de4543ccbf
- novox/mesh-host@74ed7cbe0ab4 before novox/mesh-controller@d4de4543ccbf: engine before controller
- composed together: pass — every machine composes with the change as it did without (4 of 4 compose)
**Machines**
- mesh-controller on novox: passed — healthy 3 times over 4m7s
- mesh-controller on the rest running it: sent
**Transitions**
- 2026-10-12 13:39 (new) → proposed (announced): novox/mesh-controller#263's head announced
- 2026-10-12 13:48 proposed → checked (checked): the verdict names this commit
- 2026-10-12 13:48 checked → ready (accepted): the gate passed or warned, and the repository's own check did not fail
- 2026-10-12 14:11 ready → published (merged): on the trunk its modules follow: merged there, and its walk opened — or nothing for a walk to move
- 2026-10-12 14:11 published → delivering (go): its walk started: let go by this owner in its group's order, by a person, or on the controller's own path — not a failed walk that carried a later commit for it, after which it waits for a walk of its own commit
- 2026-10-12 14:16 delivering → delivered (done): its walk is done, carrying its own commit: every machine of its deploy plan passed, or was left as its policy says
The commit's note under `refs/notes/mesh-plan` keeps every transition: `git log --notes=mesh-plan`.
build-agent keeps data (ADR 0233): what it holds is backed up before it moves
route-proxy provides route: its consumers are sent again after it
route-proxy keeps data (ADR 0233): what it holds is backed up before it moves
grants step: if this changes what the bus's user list says (a new tool, subject or user), novox is sent that list alone, every build there kept, before build-agent, route-proxy is judged on its first machine
Run by the build seat on g14 as build-1791809446101369103 (builds --log build-1791809446101369103).
--- the gate: build-agent, mesh-controller, route-proxy
mesh-controller: ok, 58 tool(s); WARNING: merge-window ask(s) for a setting and do(es) not say whether it is trusted, so it counts as trusted: set at the terminal alone; say "trusted" false where nothing trusts it (novox/hq issue 339); WARNING: controller stay(s) up and say(s) not how it is ready — judged by liveness alone, refused from 2026-11-18 (ADR 0240 rule 8)
long-running resources without health: 1
files asking for a setting without saying whether it is trusted: 1
1 manifest(s) checked. Judged against the seats this binary carries; a claim on one of the mesh's own seats is judged fully at registration, and a seat declared by a module not given here reads as unknown. Identities judged on a 6-character machine name, against the bounds of the providers given here
judging 1 module(s) changed against 4 machine(s), as the snapshot of 2026-10-12T12:51:20Z says they run
composing every machine as the mesh is
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
composing every machine with the change
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
merge gate: PASS — every machine composes with the change as it did without (4 of 4 compose)
judged by controller development build against the facts of 2026-10-12T12:51:20Z
modules the change moves:
- mesh-controller (changed)
machines:
- a machine (f35): composes
- holds mesh-dns-resolver (idt): composes
- holds git, holds mesh-artifact-store, holds mesh-broker, holds mesh-catalog, holds mesh-delivery, holds mesh-dns-resolver, holds mesh-store, holds mesh-vault, holds npm-package-registry, the control node, the hub (ygcyr): composes
- a machine (yvbpsm): composes
a merge would rebuild 3 module(s) in 3 tier(s): build-agent, mesh-controller, route-proxy
--- its merge-check.sh (go toolchain)
ok github.com/novox/mesh-controller/cmd/mesh-builder 1.034s
ok github.com/novox/mesh-controller/cmd/mesh-controller 253.889s
ok github.com/novox/mesh-controller/examples/objectstore-provisioner 1.018s
ok github.com/novox/mesh-controller/examples/postgres-provisioner 1.030s
ok github.com/novox/mesh-controller/examples/redis-provisioner 1.017s
ok github.com/novox/mesh-controller/examples/route-proxy 1.160s
ok github.com/novox/mesh-controller/internal/artifacts 1.054s
? github.com/novox/mesh-controller/internal/beside [no test files]
ok github.com/novox/mesh-controller/internal/broker 3.964s
ok github.com/novox/mesh-controller/internal/builder 29.442s
ok github.com/novox/mesh-controller/internal/catalogue 7.106s
ok github.com/novox/mesh-controller/internal/conditions 2.449s
ok github.com/novox/mesh-controller/internal/envfile 1.032s
ok github.com/novox/mesh-controller/internal/facts 1.037s
ok github.com/novox/mesh-controller/internal/identity 4.249s
ok github.com/novox/mesh-controller/internal/inventory 85.668s
ok github.com/novox/mesh-controller/internal/lease 17.702s
ok github.com/novox/mesh-controller/internal/licences 3.867s
ok github.com/novox/mesh-controller/internal/licences/adapters 1.012s
ok github.com/novox/mesh-controller/internal/link 58.867s
ok github.com/novox/mesh-controller/internal/lint 2.231s
ok github.com/novox/mesh-controller/internal/outward 1.038s
ok github.com/novox/mesh-controller/internal/overlay 1.031s
ok github.com/novox/mesh-controller/internal/secrets 1.457s
ok github.com/novox/mesh-controller/internal/store 3.346s
ok github.com/novox/mesh-controller/internal/testbus 1.181s
ok github.com/novox/mesh-controller/internal/token 1.018s
store tests, each run against the store:
./cmd/mesh-controller --- PASS: TestAStopWithoutAMergeWalksTheOthersAgain (0.37s)
./cmd/mesh-controller --- PASS: TestAStopWithoutAMergeEndsTheMergesThatContainIt (0.35s)
./cmd/mesh-controller --- PASS: TestAStopWalksTheOthersAgainInTheOpenBatch (0.35s)
./cmd/mesh-controller --- PASS: TestAStopNamesWhatStaysOnTheMachinesAndWhatStillCarriesIt (0.35s)
./cmd/mesh-controller --- PASS: TestAStopWithoutAMergeLeavesItOutOfABatch (0.39s)
./cmd/mesh-controller --- PASS: TestSavingPlansAndMovingMergesIsOneAct (0.30s)
./cmd/mesh-controller --- PASS: TestNodeShowAndStatusReadTheLastSendFromTheStore (0.38s)
./cmd/mesh-controller --- PASS: TestTheConditionsVerbReadsTheLooksTheBucketNoLongerHolds (0.24s)
./cmd/mesh-controller --- PASS: TestAWalkLetGoAndStoppedByVerbsIsSaidStartedAndEnded (0.38s)
./cmd/mesh-controller --- PASS: TestAWalkStoppedOrClosedByAPersonIsSaidEndedStopped (0.38s)
./cmd/mesh-controller --- PASS: TestAWalkRetriedByAPersonIsSaidStartedAgain (0.35s)
./cmd/mesh-controller --- PASS: TestABacklogReleasedByAPersonIsSaidStarted (0.36s)
./cmd/mesh-controller --- PASS: TestAWalkEndedByCancelOrClearIsSaidEnded (0.41s)
./cmd/mesh-controller --- PASS: TestAWalkRestartedByARebuildIsSaidStarted (0.35s)
./cmd/mesh-controller --- PASS: TestAWalkStoppedByHandAndRetriedEndsFailedWhenItFailsAgain (0.37s)
./internal/inventory --- PASS: TestTheLastSendSaysWhatTheMachineAnswered (0.21s)
./internal/inventory --- PASS: TestAReportIsTheLastSendsAnswerOnlyWhenItIsAboutThatSend (0.20s)
./internal/inventory --- PASS: TestAnIssuedValueIsSealedToEveryHolderOnceAndNotAgain (0.21s)
./internal/inventory --- PASS: TestAnIssuedValueIsSealedAgainAfterAKeyChangeOrAnAcceptance (0.21s)
./internal/inventory --- PASS: TestAConditionsLooksAreKeptBesideTheBucket (0.20s)
./internal/inventory --- PASS: TestAWalksStartIsKeptAndARefusedSaveLeavesItUnsaid (0.19s)
**Merge check** at `3e10ad15`
- `mesh/merge-gate`: **PASS** — every machine composes with the change as it did without (4 of 4 compose) (modules: build-agent, mesh-controller, route-proxy)
- `mesh/repo-check`: **PASS** — its merge-check.sh passed
**Delivery plan** — builds build-agent, mesh-controller, route-proxy → ace, g14, novox, shanks; no bus step
- tier 0: mesh-controller
- tier 1: build-agent
- tier 2: route-proxy
- ace: receives build-agent, route-proxy
- g14: receives build-agent
- novox: receives mesh-controller, build-agent, route-proxy
- shanks: receives build-agent
- build-agent keeps data (ADR 0233): what it holds is backed up before it moves
- route-proxy provides route: its consumers are sent again after it
- route-proxy keeps data (ADR 0233): what it holds is backed up before it moves
- grants step: if this changes what the bus's user list says (a new tool, subject or user), novox is sent that list alone, every build there kept, before build-agent, route-proxy is judged on its first machine
Run by the build seat on g14 as `build-1791809446101369103` (`builds --log build-1791809446101369103`).
```
--- the gate: build-agent, mesh-controller, route-proxy
mesh-controller: ok, 58 tool(s); WARNING: merge-window ask(s) for a setting and do(es) not say whether it is trusted, so it counts as trusted: set at the terminal alone; say "trusted" false where nothing trusts it (novox/hq issue 339); WARNING: controller stay(s) up and say(s) not how it is ready — judged by liveness alone, refused from 2026-11-18 (ADR 0240 rule 8)
long-running resources without health: 1
files asking for a setting without saying whether it is trusted: 1
1 manifest(s) checked. Judged against the seats this binary carries; a claim on one of the mesh's own seats is judged fully at registration, and a seat declared by a module not given here reads as unknown. Identities judged on a 6-character machine name, against the bounds of the providers given here
judging 1 module(s) changed against 4 machine(s), as the snapshot of 2026-10-12T12:51:20Z says they run
composing every machine as the mesh is
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
composing every machine with the change
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
merge gate: PASS — every machine composes with the change as it did without (4 of 4 compose)
judged by controller development build against the facts of 2026-10-12T12:51:20Z
modules the change moves:
- mesh-controller (changed)
machines:
- a machine (f35): composes
- holds mesh-dns-resolver (idt): composes
- holds git, holds mesh-artifact-store, holds mesh-broker, holds mesh-catalog, holds mesh-delivery, holds mesh-dns-resolver, holds mesh-store, holds mesh-vault, holds npm-package-registry, the control node, the hub (ygcyr): composes
- a machine (yvbpsm): composes
a merge would rebuild 3 module(s) in 3 tier(s): build-agent, mesh-controller, route-proxy
--- its merge-check.sh (go toolchain)
ok github.com/novox/mesh-controller/cmd/mesh-builder 1.034s
ok github.com/novox/mesh-controller/cmd/mesh-controller 253.889s
ok github.com/novox/mesh-controller/examples/objectstore-provisioner 1.018s
ok github.com/novox/mesh-controller/examples/postgres-provisioner 1.030s
ok github.com/novox/mesh-controller/examples/redis-provisioner 1.017s
ok github.com/novox/mesh-controller/examples/route-proxy 1.160s
ok github.com/novox/mesh-controller/internal/artifacts 1.054s
? github.com/novox/mesh-controller/internal/beside [no test files]
ok github.com/novox/mesh-controller/internal/broker 3.964s
ok github.com/novox/mesh-controller/internal/builder 29.442s
ok github.com/novox/mesh-controller/internal/catalogue 7.106s
ok github.com/novox/mesh-controller/internal/conditions 2.449s
ok github.com/novox/mesh-controller/internal/envfile 1.032s
ok github.com/novox/mesh-controller/internal/facts 1.037s
ok github.com/novox/mesh-controller/internal/identity 4.249s
ok github.com/novox/mesh-controller/internal/inventory 85.668s
ok github.com/novox/mesh-controller/internal/lease 17.702s
ok github.com/novox/mesh-controller/internal/licences 3.867s
ok github.com/novox/mesh-controller/internal/licences/adapters 1.012s
ok github.com/novox/mesh-controller/internal/link 58.867s
ok github.com/novox/mesh-controller/internal/lint 2.231s
ok github.com/novox/mesh-controller/internal/outward 1.038s
ok github.com/novox/mesh-controller/internal/overlay 1.031s
ok github.com/novox/mesh-controller/internal/secrets 1.457s
ok github.com/novox/mesh-controller/internal/store 3.346s
ok github.com/novox/mesh-controller/internal/testbus 1.181s
ok github.com/novox/mesh-controller/internal/token 1.018s
store tests, each run against the store:
./cmd/mesh-controller --- PASS: TestAStopWithoutAMergeWalksTheOthersAgain (0.37s)
./cmd/mesh-controller --- PASS: TestAStopWithoutAMergeEndsTheMergesThatContainIt (0.35s)
./cmd/mesh-controller --- PASS: TestAStopWalksTheOthersAgainInTheOpenBatch (0.35s)
./cmd/mesh-controller --- PASS: TestAStopNamesWhatStaysOnTheMachinesAndWhatStillCarriesIt (0.35s)
./cmd/mesh-controller --- PASS: TestAStopWithoutAMergeLeavesItOutOfABatch (0.39s)
./cmd/mesh-controller --- PASS: TestSavingPlansAndMovingMergesIsOneAct (0.30s)
./cmd/mesh-controller --- PASS: TestNodeShowAndStatusReadTheLastSendFromTheStore (0.38s)
./cmd/mesh-controller --- PASS: TestTheConditionsVerbReadsTheLooksTheBucketNoLongerHolds (0.24s)
./cmd/mesh-controller --- PASS: TestAWalkLetGoAndStoppedByVerbsIsSaidStartedAndEnded (0.38s)
./cmd/mesh-controller --- PASS: TestAWalkStoppedOrClosedByAPersonIsSaidEndedStopped (0.38s)
./cmd/mesh-controller --- PASS: TestAWalkRetriedByAPersonIsSaidStartedAgain (0.35s)
./cmd/mesh-controller --- PASS: TestABacklogReleasedByAPersonIsSaidStarted (0.36s)
./cmd/mesh-controller --- PASS: TestAWalkEndedByCancelOrClearIsSaidEnded (0.41s)
./cmd/mesh-controller --- PASS: TestAWalkRestartedByARebuildIsSaidStarted (0.35s)
./cmd/mesh-controller --- PASS: TestAWalkStoppedByHandAndRetriedEndsFailedWhenItFailsAgain (0.37s)
./internal/inventory --- PASS: TestTheLastSendSaysWhatTheMachineAnswered (0.21s)
./internal/inventory --- PASS: TestAReportIsTheLastSendsAnswerOnlyWhenItIsAboutThatSend (0.20s)
./internal/inventory --- PASS: TestAnIssuedValueIsSealedToEveryHolderOnceAndNotAgain (0.21s)
./internal/inventory --- PASS: TestAnIssuedValueIsSealedAgainAfterAKeyChangeOrAnAcceptance (0.21s)
./internal/inventory --- PASS: TestAConditionsLooksAreKeptBesideTheBucket (0.20s)
./internal/inventory --- PASS: TestAWalksStartIsKeptAndARefusedSaveLeavesItUnsaid (0.19s)
```
mesh/merge-gate: PASS — every machine composes with the change as it did without (4 of 4 compose) (modules: build-agent, mesh-controller)
mesh/repo-check: PASS — its merge-check.sh passed
Delivery plan — builds build-agent, mesh-controller → ace, g14, novox, shanks; no bus step
tier 0: mesh-controller
tier 1: build-agent
ace: receives build-agent
g14: receives build-agent
novox: receives mesh-controller, build-agent
shanks: receives build-agent
build-agent keeps data (ADR 0233): what it holds is backed up before it moves
grants step: if this changes what the bus's user list says (a new tool, subject or user), novox is sent that list alone, every build there kept, before build-agent is judged on its first machine
Run by the build seat on g14 as build-1791810347306736912 (builds --log build-1791810347306736912).
--- the gate: build-agent, mesh-controller
mesh-controller: ok, 58 tool(s); WARNING: merge-window ask(s) for a setting and do(es) not say whether it is trusted, so it counts as trusted: set at the terminal alone; say "trusted" false where nothing trusts it (novox/hq issue 339); WARNING: controller stay(s) up and say(s) not how it is ready — judged by liveness alone, refused from 2026-11-18 (ADR 0240 rule 8)
long-running resources without health: 1
files asking for a setting without saying whether it is trusted: 1
1 manifest(s) checked. Judged against the seats this binary carries; a claim on one of the mesh's own seats is judged fully at registration, and a seat declared by a module not given here reads as unknown. Identities judged on a 6-character machine name, against the bounds of the providers given here
judging 1 module(s) changed against 4 machine(s), as the snapshot of 2026-10-12T13:06:42Z says they run
composing every machine as the mesh is
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
composing every machine with the change
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
merge gate: PASS — every machine composes with the change as it did without (4 of 4 compose)
judged by controller development build against the facts of 2026-10-12T13:06:42Z
modules the change moves:
- mesh-controller (changed)
machines:
- a machine (f35): composes
- holds mesh-dns-resolver (idt): composes
- holds git, holds mesh-artifact-store, holds mesh-broker, holds mesh-catalog, holds mesh-delivery, holds mesh-dns-resolver, holds mesh-store, holds mesh-vault, holds npm-package-registry, the control node, the hub (ygcyr): composes
- a machine (yvbpsm): composes
a merge would rebuild 2 module(s) in 2 tier(s): build-agent, mesh-controller
--- its merge-check.sh (go toolchain)
ok github.com/novox/mesh-controller/cmd/mesh-builder 1.047s
ok github.com/novox/mesh-controller/cmd/mesh-controller 253.937s
ok github.com/novox/mesh-controller/examples/objectstore-provisioner 1.015s
ok github.com/novox/mesh-controller/examples/postgres-provisioner 1.024s
ok github.com/novox/mesh-controller/examples/redis-provisioner 1.013s
ok github.com/novox/mesh-controller/examples/route-proxy 1.151s
ok github.com/novox/mesh-controller/internal/artifacts 1.044s
? github.com/novox/mesh-controller/internal/beside [no test files]
ok github.com/novox/mesh-controller/internal/broker 4.067s
ok github.com/novox/mesh-controller/internal/builder 30.315s
ok github.com/novox/mesh-controller/internal/catalogue 6.934s
ok github.com/novox/mesh-controller/internal/conditions 2.444s
ok github.com/novox/mesh-controller/internal/envfile 1.024s
ok github.com/novox/mesh-controller/internal/facts 1.025s
ok github.com/novox/mesh-controller/internal/identity 5.192s
ok github.com/novox/mesh-controller/internal/inventory 88.299s
ok github.com/novox/mesh-controller/internal/lease 17.700s
ok github.com/novox/mesh-controller/internal/licences 4.801s
ok github.com/novox/mesh-controller/internal/licences/adapters 1.010s
ok github.com/novox/mesh-controller/internal/link 61.808s
ok github.com/novox/mesh-controller/internal/lint 2.154s
ok github.com/novox/mesh-controller/internal/outward 1.040s
ok github.com/novox/mesh-controller/internal/overlay 1.028s
ok github.com/novox/mesh-controller/internal/secrets 1.455s
ok github.com/novox/mesh-controller/internal/store 4.232s
ok github.com/novox/mesh-controller/internal/testbus 1.181s
ok github.com/novox/mesh-controller/internal/token 1.019s
store tests, each run against the store:
./cmd/mesh-controller --- PASS: TestAStopWithoutAMergeWalksTheOthersAgain (0.34s)
./cmd/mesh-controller --- PASS: TestAStopWithoutAMergeEndsTheMergesThatContainIt (0.35s)
./cmd/mesh-controller --- PASS: TestAStopWalksTheOthersAgainInTheOpenBatch (0.35s)
./cmd/mesh-controller --- PASS: TestAStopNamesWhatStaysOnTheMachinesAndWhatStillCarriesIt (0.35s)
./cmd/mesh-controller --- PASS: TestAStopWithoutAMergeLeavesItOutOfABatch (0.33s)
./cmd/mesh-controller --- PASS: TestSavingPlansAndMovingMergesIsOneAct (0.30s)
./cmd/mesh-controller --- PASS: TestNodeShowAndStatusReadTheLastSendFromTheStore (0.39s)
./cmd/mesh-controller --- PASS: TestTheConditionsVerbReadsTheLooksTheBucketNoLongerHolds (0.25s)
./cmd/mesh-controller --- PASS: TestAWalkLetGoAndStoppedByVerbsIsSaidStartedAndEnded (0.39s)
./cmd/mesh-controller --- PASS: TestAWalkStoppedOrClosedByAPersonIsSaidEndedStopped (0.39s)
./cmd/mesh-controller --- PASS: TestAWalkRetriedByAPersonIsSaidStartedAgain (0.38s)
./cmd/mesh-controller --- PASS: TestABacklogReleasedByAPersonIsSaidStarted (0.35s)
./cmd/mesh-controller --- PASS: TestAWalkEndedByCancelOrClearIsSaidEnded (0.40s)
./cmd/mesh-controller --- PASS: TestAWalkRestartedByARebuildIsSaidStarted (0.35s)
./cmd/mesh-controller --- PASS: TestAWalkStoppedByHandAndRetriedEndsFailedWhenItFailsAgain (0.37s)
./internal/inventory --- PASS: TestTheLastSendSaysWhatTheMachineAnswered (0.20s)
./internal/inventory --- PASS: TestAReportIsTheLastSendsAnswerOnlyWhenItIsAboutThatSend (0.21s)
./internal/inventory --- PASS: TestAnIssuedValueIsSealedToEveryHolderOnceAndNotAgain (0.21s)
./internal/inventory --- PASS: TestAnIssuedValueIsSealedAgainAfterAKeyChangeOrAnAcceptance (0.20s)
./internal/inventory --- PASS: TestAConditionsLooksAreKeptBesideTheBucket (0.21s)
./internal/inventory --- PASS: TestAWalksStartIsKeptAndARefusedSaveLeavesItUnsaid (0.20s)
**Merge check** at `3e10ad15`
- `mesh/merge-gate`: **PASS** — every machine composes with the change as it did without (4 of 4 compose) (modules: build-agent, mesh-controller)
- `mesh/repo-check`: **PASS** — its merge-check.sh passed
**Delivery plan** — builds build-agent, mesh-controller → ace, g14, novox, shanks; no bus step
- tier 0: mesh-controller
- tier 1: build-agent
- ace: receives build-agent
- g14: receives build-agent
- novox: receives mesh-controller, build-agent
- shanks: receives build-agent
- build-agent keeps data (ADR 0233): what it holds is backed up before it moves
- grants step: if this changes what the bus's user list says (a new tool, subject or user), novox is sent that list alone, every build there kept, before build-agent is judged on its first machine
Run by the build seat on g14 as `build-1791810347306736912` (`builds --log build-1791810347306736912`).
```
--- the gate: build-agent, mesh-controller
mesh-controller: ok, 58 tool(s); WARNING: merge-window ask(s) for a setting and do(es) not say whether it is trusted, so it counts as trusted: set at the terminal alone; say "trusted" false where nothing trusts it (novox/hq issue 339); WARNING: controller stay(s) up and say(s) not how it is ready — judged by liveness alone, refused from 2026-11-18 (ADR 0240 rule 8)
long-running resources without health: 1
files asking for a setting without saying whether it is trusted: 1
1 manifest(s) checked. Judged against the seats this binary carries; a claim on one of the mesh's own seats is judged fully at registration, and a seat declared by a module not given here reads as unknown. Identities judged on a 6-character machine name, against the bounds of the providers given here
judging 1 module(s) changed against 4 machine(s), as the snapshot of 2026-10-12T13:06:42Z says they run
composing every machine as the mesh is
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
composing every machine with the change
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
merge gate: PASS — every machine composes with the change as it did without (4 of 4 compose)
judged by controller development build against the facts of 2026-10-12T13:06:42Z
modules the change moves:
- mesh-controller (changed)
machines:
- a machine (f35): composes
- holds mesh-dns-resolver (idt): composes
- holds git, holds mesh-artifact-store, holds mesh-broker, holds mesh-catalog, holds mesh-delivery, holds mesh-dns-resolver, holds mesh-store, holds mesh-vault, holds npm-package-registry, the control node, the hub (ygcyr): composes
- a machine (yvbpsm): composes
a merge would rebuild 2 module(s) in 2 tier(s): build-agent, mesh-controller
--- its merge-check.sh (go toolchain)
ok github.com/novox/mesh-controller/cmd/mesh-builder 1.047s
ok github.com/novox/mesh-controller/cmd/mesh-controller 253.937s
ok github.com/novox/mesh-controller/examples/objectstore-provisioner 1.015s
ok github.com/novox/mesh-controller/examples/postgres-provisioner 1.024s
ok github.com/novox/mesh-controller/examples/redis-provisioner 1.013s
ok github.com/novox/mesh-controller/examples/route-proxy 1.151s
ok github.com/novox/mesh-controller/internal/artifacts 1.044s
? github.com/novox/mesh-controller/internal/beside [no test files]
ok github.com/novox/mesh-controller/internal/broker 4.067s
ok github.com/novox/mesh-controller/internal/builder 30.315s
ok github.com/novox/mesh-controller/internal/catalogue 6.934s
ok github.com/novox/mesh-controller/internal/conditions 2.444s
ok github.com/novox/mesh-controller/internal/envfile 1.024s
ok github.com/novox/mesh-controller/internal/facts 1.025s
ok github.com/novox/mesh-controller/internal/identity 5.192s
ok github.com/novox/mesh-controller/internal/inventory 88.299s
ok github.com/novox/mesh-controller/internal/lease 17.700s
ok github.com/novox/mesh-controller/internal/licences 4.801s
ok github.com/novox/mesh-controller/internal/licences/adapters 1.010s
ok github.com/novox/mesh-controller/internal/link 61.808s
ok github.com/novox/mesh-controller/internal/lint 2.154s
ok github.com/novox/mesh-controller/internal/outward 1.040s
ok github.com/novox/mesh-controller/internal/overlay 1.028s
ok github.com/novox/mesh-controller/internal/secrets 1.455s
ok github.com/novox/mesh-controller/internal/store 4.232s
ok github.com/novox/mesh-controller/internal/testbus 1.181s
ok github.com/novox/mesh-controller/internal/token 1.019s
store tests, each run against the store:
./cmd/mesh-controller --- PASS: TestAStopWithoutAMergeWalksTheOthersAgain (0.34s)
./cmd/mesh-controller --- PASS: TestAStopWithoutAMergeEndsTheMergesThatContainIt (0.35s)
./cmd/mesh-controller --- PASS: TestAStopWalksTheOthersAgainInTheOpenBatch (0.35s)
./cmd/mesh-controller --- PASS: TestAStopNamesWhatStaysOnTheMachinesAndWhatStillCarriesIt (0.35s)
./cmd/mesh-controller --- PASS: TestAStopWithoutAMergeLeavesItOutOfABatch (0.33s)
./cmd/mesh-controller --- PASS: TestSavingPlansAndMovingMergesIsOneAct (0.30s)
./cmd/mesh-controller --- PASS: TestNodeShowAndStatusReadTheLastSendFromTheStore (0.39s)
./cmd/mesh-controller --- PASS: TestTheConditionsVerbReadsTheLooksTheBucketNoLongerHolds (0.25s)
./cmd/mesh-controller --- PASS: TestAWalkLetGoAndStoppedByVerbsIsSaidStartedAndEnded (0.39s)
./cmd/mesh-controller --- PASS: TestAWalkStoppedOrClosedByAPersonIsSaidEndedStopped (0.39s)
./cmd/mesh-controller --- PASS: TestAWalkRetriedByAPersonIsSaidStartedAgain (0.38s)
./cmd/mesh-controller --- PASS: TestABacklogReleasedByAPersonIsSaidStarted (0.35s)
./cmd/mesh-controller --- PASS: TestAWalkEndedByCancelOrClearIsSaidEnded (0.40s)
./cmd/mesh-controller --- PASS: TestAWalkRestartedByARebuildIsSaidStarted (0.35s)
./cmd/mesh-controller --- PASS: TestAWalkStoppedByHandAndRetriedEndsFailedWhenItFailsAgain (0.37s)
./internal/inventory --- PASS: TestTheLastSendSaysWhatTheMachineAnswered (0.20s)
./internal/inventory --- PASS: TestAReportIsTheLastSendsAnswerOnlyWhenItIsAboutThatSend (0.21s)
./internal/inventory --- PASS: TestAnIssuedValueIsSealedToEveryHolderOnceAndNotAgain (0.21s)
./internal/inventory --- PASS: TestAnIssuedValueIsSealedAgainAfterAKeyChangeOrAnAcceptance (0.20s)
./internal/inventory --- PASS: TestAConditionsLooksAreKeptBesideTheBucket (0.21s)
./internal/inventory --- PASS: TestAWalksStartIsKeptAndARefusedSaveLeavesItUnsaid (0.20s)
```
Independent review of the issue 233 delivery group (this and mesh-host #90), at 3e10ad1.
What I ran.go vet ./... is clean. merge-check.sh passes against a throwaway Postgres (MESH_TEST_POSTGRES set): gofmt, vet, go test -race and the listed store tests all PASS. All four checks are green at this head: merge-gate, repo-check, delivery and delivery-group.
What holds.
S21 is a warning keyed machine.<m>.apply-failing, one per machine, so it does not flood.
It is raised past one apply interval over at least 2 reports, and suppressed while the machine is asleep or silent past its heartbeat bound.
The headline " cannot install packages" appears when the the package manager ( key is present. The path stays out of the summary and the probe's words go into the evidence.
The suppression row and the S21 tests are sound against the inputs they are given.
One side note: S21 fires for any persisting partial apply, not only the package manager. That matches the decision's (3).
Blocking: S21's inputs do not arrive as it assumes, so it neither counts nor clears in the live flow.Times advances only when a report arrives. The node-engine's reconcile is said unasked only when worthSaying && adoption fingerprint differs, and that fingerprint carries neither Refused nor Failed (mesh-host cmd/mesh-host/main.go, adoptionFingerprint and reconcileKept). Two results follow:
A failure that persists with no new delivery stays at Times = 1, so S21 is never raised.
A machine that recovers on a reconcile is never said to be clean. In the 233 replay the restoring delivery's own report is partial, because the probe runs before the file is written, and the clean reconcile after it is silent. S21, once raised, stays raised over a healthy machine until the next push.
"Cleared by a clean apply" therefore holds only for a clean delivery. The fix needs mesh-host #90 to say a reconcile whose outcome changed, and either:
to say each failing reconcile, so Times ≥ 2 works as written; or
to say only changes, with S21 judging by time since the first report while the machine is heard.
A test should prove the clear on a reconcile.
Verdict: request changes at 3e10ad1
Independent review of the issue 233 delivery group (this and mesh-host #90), at 3e10ad1.
**What I ran.** `go vet ./...` is clean. `merge-check.sh` passes against a throwaway Postgres (MESH_TEST_POSTGRES set): gofmt, vet, `go test -race` and the listed store tests all PASS. All four checks are green at this head: merge-gate, repo-check, delivery and delivery-group.
**What holds.**
- S21 is a warning keyed `machine.<m>.apply-failing`, one per machine, so it does not flood.
- It is raised past one apply interval over at least 2 reports, and suppressed while the machine is asleep or silent past its heartbeat bound.
- The headline "<m> cannot install packages" appears when the `the package manager (` key is present. The path stays out of the summary and the probe's words go into the evidence.
- The suppression row and the S21 tests are sound against the inputs they are given.
- One side note: S21 fires for any persisting partial apply, not only the package manager. That matches the decision's (3).
**Blocking: S21's inputs do not arrive as it assumes, so it neither counts nor clears in the live flow.** `Times` advances only when a report arrives. The node-engine's reconcile is said unasked only when `worthSaying && adoption fingerprint differs`, and that fingerprint carries neither Refused nor Failed (mesh-host cmd/mesh-host/main.go, adoptionFingerprint and reconcileKept). Two results follow:
- A failure that persists with no new delivery stays at Times = 1, so S21 is never raised.
- A machine that recovers on a reconcile is never said to be clean. In the 233 replay the restoring delivery's own report is partial, because the probe runs before the file is written, and the clean reconcile after it is silent. S21, once raised, stays raised over a healthy machine until the next push.
"Cleared by a clean apply" therefore holds only for a clean delivery. The fix needs mesh-host #90 to say a reconcile whose outcome changed, and either:
- to say each failing reconcile, so `Times ≥ 2` works as written; or
- to say only changes, with S21 judging by time since the first report while the machine is heard.
A test should prove the clear on a reconcile.
mesh/merge-gate: PASS — every machine composes with the change as it did without (4 of 4 compose) (modules: mesh-controller)
mesh/repo-check: PASS — its merge-check.sh passed
Delivery plan — builds mesh-controller → novox; no bus step
tier 0: mesh-controller
novox: receives mesh-controller
read by no module's build: .gitignore, cmd/mesh-controller/apply_failing_test.go, cmd/mesh-controller/signals_test.go
Run by the build seat on novox as build-1791812356379633013 (builds --log build-1791812356379633013).
--- the gate: mesh-controller
mesh-controller: ok, 58 tool(s); WARNING: merge-window ask(s) for a setting and do(es) not say whether it is trusted, so it counts as trusted: set at the terminal alone; say "trusted" false where nothing trusts it (novox/hq issue 339); WARNING: controller stay(s) up and say(s) not how it is ready — judged by liveness alone, refused from 2026-11-18 (ADR 0240 rule 8)
long-running resources without health: 1
files asking for a setting without saying whether it is trusted: 1
1 manifest(s) checked. Judged against the seats this binary carries; a claim on one of the mesh's own seats is judged fully at registration, and a seat declared by a module not given here reads as unknown. Identities judged on a 6-character machine name, against the bounds of the providers given here
judging 1 module(s) changed against 4 machine(s), as the snapshot of 2026-10-12T13:33:41Z says they run
composing every machine as the mesh is
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
composing every machine with the change
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
merge gate: PASS — every machine composes with the change as it did without (4 of 4 compose)
judged by controller development build against the facts of 2026-10-12T13:33:41Z
modules the change moves:
- mesh-controller (changed)
machines:
- a machine (f35): composes
- holds mesh-dns-resolver (idt): composes
- holds git, holds mesh-artifact-store, holds mesh-broker, holds mesh-catalog, holds mesh-delivery, holds mesh-dns-resolver, holds mesh-store, holds mesh-vault, holds npm-package-registry, the control node, the hub (ygcyr): composes
- a machine (yvbpsm): composes
a merge would rebuild 1 module(s) in 1 tier(s): mesh-controller
notes:
- .gitignore, cmd/mesh-controller/apply_failing_test.go and cmd/mesh-controller/signals_test.go read by no module's build, and rebuild nothing (issue 280)
--- its merge-check.sh (go toolchain)
ok github.com/novox/mesh-controller/cmd/mesh-builder 1.083s
ok github.com/novox/mesh-controller/cmd/mesh-controller 329.182s
ok github.com/novox/mesh-controller/examples/objectstore-provisioner 1.016s
ok github.com/novox/mesh-controller/examples/postgres-provisioner 1.025s
ok github.com/novox/mesh-controller/examples/redis-provisioner 1.016s
ok github.com/novox/mesh-controller/examples/route-proxy 1.126s
ok github.com/novox/mesh-controller/internal/artifacts 1.063s
? github.com/novox/mesh-controller/internal/beside [no test files]
ok github.com/novox/mesh-controller/internal/broker 5.724s
ok github.com/novox/mesh-controller/internal/builder 55.075s
ok github.com/novox/mesh-controller/internal/catalogue 8.885s
ok github.com/novox/mesh-controller/internal/conditions 2.714s
ok github.com/novox/mesh-controller/internal/envfile 1.030s
ok github.com/novox/mesh-controller/internal/facts 1.033s
ok github.com/novox/mesh-controller/internal/identity 11.950s
ok github.com/novox/mesh-controller/internal/inventory 141.862s
ok github.com/novox/mesh-controller/internal/lease 18.112s
ok github.com/novox/mesh-controller/internal/licences 11.205s
ok github.com/novox/mesh-controller/internal/licences/adapters 1.018s
ok github.com/novox/mesh-controller/internal/link 78.705s
ok github.com/novox/mesh-controller/internal/lint 2.667s
ok github.com/novox/mesh-controller/internal/outward 1.050s
ok github.com/novox/mesh-controller/internal/overlay 1.030s
ok github.com/novox/mesh-controller/internal/secrets 1.463s
ok github.com/novox/mesh-controller/internal/store 7.532s
ok github.com/novox/mesh-controller/internal/testbus 1.246s
ok github.com/novox/mesh-controller/internal/token 1.022s
store tests, each run against the store:
./cmd/mesh-controller --- PASS: TestAStopWithoutAMergeWalksTheOthersAgain (0.57s)
./cmd/mesh-controller --- PASS: TestAStopWithoutAMergeEndsTheMergesThatContainIt (0.51s)
./cmd/mesh-controller --- PASS: TestAStopWalksTheOthersAgainInTheOpenBatch (0.57s)
./cmd/mesh-controller --- PASS: TestAStopNamesWhatStaysOnTheMachinesAndWhatStillCarriesIt (0.52s)
./cmd/mesh-controller --- PASS: TestAStopWithoutAMergeLeavesItOutOfABatch (0.47s)
./cmd/mesh-controller --- PASS: TestSavingPlansAndMovingMergesIsOneAct (0.48s)
./cmd/mesh-controller --- PASS: TestNodeShowAndStatusReadTheLastSendFromTheStore (0.53s)
./cmd/mesh-controller --- PASS: TestTheConditionsVerbReadsTheLooksTheBucketNoLongerHolds (0.56s)
./cmd/mesh-controller --- PASS: TestAWalkLetGoAndStoppedByVerbsIsSaidStartedAndEnded (0.52s)
./cmd/mesh-controller --- PASS: TestAWalkStoppedOrClosedByAPersonIsSaidEndedStopped (0.50s)
./cmd/mesh-controller --- PASS: TestAWalkRetriedByAPersonIsSaidStartedAgain (0.51s)
./cmd/mesh-controller --- PASS: TestABacklogReleasedByAPersonIsSaidStarted (0.52s)
./cmd/mesh-controller --- PASS: TestAWalkEndedByCancelOrClearIsSaidEnded (0.54s)
./cmd/mesh-controller --- PASS: TestAWalkRestartedByARebuildIsSaidStarted (0.43s)
./cmd/mesh-controller --- PASS: TestAWalkStoppedByHandAndRetriedEndsFailedWhenItFailsAgain (0.53s)
./internal/inventory --- PASS: TestTheLastSendSaysWhatTheMachineAnswered (0.28s)
./internal/inventory --- PASS: TestAReportIsTheLastSendsAnswerOnlyWhenItIsAboutThatSend (0.28s)
./internal/inventory --- PASS: TestAnIssuedValueIsSealedToEveryHolderOnceAndNotAgain (0.30s)
./internal/inventory --- PASS: TestAnIssuedValueIsSealedAgainAfterAKeyChangeOrAnAcceptance (0.29s)
./internal/inventory --- PASS: TestAConditionsLooksAreKeptBesideTheBucket (0.28s)
./internal/inventory --- PASS: TestAWalksStartIsKeptAndARefusedSaveLeavesItUnsaid (0.27s)
**Merge check** at `d4de4543`
- `mesh/merge-gate`: **PASS** — every machine composes with the change as it did without (4 of 4 compose) (modules: mesh-controller)
- `mesh/repo-check`: **PASS** — its merge-check.sh passed
**Delivery plan** — builds mesh-controller → novox; no bus step
- tier 0: mesh-controller
- novox: receives mesh-controller
- read by no module's build: .gitignore, cmd/mesh-controller/apply_failing_test.go, cmd/mesh-controller/signals_test.go
Run by the build seat on novox as `build-1791812356379633013` (`builds --log build-1791812356379633013`).
```
--- the gate: mesh-controller
mesh-controller: ok, 58 tool(s); WARNING: merge-window ask(s) for a setting and do(es) not say whether it is trusted, so it counts as trusted: set at the terminal alone; say "trusted" false where nothing trusts it (novox/hq issue 339); WARNING: controller stay(s) up and say(s) not how it is ready — judged by liveness alone, refused from 2026-11-18 (ADR 0240 rule 8)
long-running resources without health: 1
files asking for a setting without saying whether it is trusted: 1
1 manifest(s) checked. Judged against the seats this binary carries; a claim on one of the mesh's own seats is judged fully at registration, and a seat declared by a module not given here reads as unknown. Identities judged on a 6-character machine name, against the bounds of the providers given here
judging 1 module(s) changed against 4 machine(s), as the snapshot of 2026-10-12T13:33:41Z says they run
composing every machine as the mesh is
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
composing every machine with the change
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
merge gate: PASS — every machine composes with the change as it did without (4 of 4 compose)
judged by controller development build against the facts of 2026-10-12T13:33:41Z
modules the change moves:
- mesh-controller (changed)
machines:
- a machine (f35): composes
- holds mesh-dns-resolver (idt): composes
- holds git, holds mesh-artifact-store, holds mesh-broker, holds mesh-catalog, holds mesh-delivery, holds mesh-dns-resolver, holds mesh-store, holds mesh-vault, holds npm-package-registry, the control node, the hub (ygcyr): composes
- a machine (yvbpsm): composes
a merge would rebuild 1 module(s) in 1 tier(s): mesh-controller
notes:
- .gitignore, cmd/mesh-controller/apply_failing_test.go and cmd/mesh-controller/signals_test.go read by no module's build, and rebuild nothing (issue 280)
--- its merge-check.sh (go toolchain)
ok github.com/novox/mesh-controller/cmd/mesh-builder 1.083s
ok github.com/novox/mesh-controller/cmd/mesh-controller 329.182s
ok github.com/novox/mesh-controller/examples/objectstore-provisioner 1.016s
ok github.com/novox/mesh-controller/examples/postgres-provisioner 1.025s
ok github.com/novox/mesh-controller/examples/redis-provisioner 1.016s
ok github.com/novox/mesh-controller/examples/route-proxy 1.126s
ok github.com/novox/mesh-controller/internal/artifacts 1.063s
? github.com/novox/mesh-controller/internal/beside [no test files]
ok github.com/novox/mesh-controller/internal/broker 5.724s
ok github.com/novox/mesh-controller/internal/builder 55.075s
ok github.com/novox/mesh-controller/internal/catalogue 8.885s
ok github.com/novox/mesh-controller/internal/conditions 2.714s
ok github.com/novox/mesh-controller/internal/envfile 1.030s
ok github.com/novox/mesh-controller/internal/facts 1.033s
ok github.com/novox/mesh-controller/internal/identity 11.950s
ok github.com/novox/mesh-controller/internal/inventory 141.862s
ok github.com/novox/mesh-controller/internal/lease 18.112s
ok github.com/novox/mesh-controller/internal/licences 11.205s
ok github.com/novox/mesh-controller/internal/licences/adapters 1.018s
ok github.com/novox/mesh-controller/internal/link 78.705s
ok github.com/novox/mesh-controller/internal/lint 2.667s
ok github.com/novox/mesh-controller/internal/outward 1.050s
ok github.com/novox/mesh-controller/internal/overlay 1.030s
ok github.com/novox/mesh-controller/internal/secrets 1.463s
ok github.com/novox/mesh-controller/internal/store 7.532s
ok github.com/novox/mesh-controller/internal/testbus 1.246s
ok github.com/novox/mesh-controller/internal/token 1.022s
store tests, each run against the store:
./cmd/mesh-controller --- PASS: TestAStopWithoutAMergeWalksTheOthersAgain (0.57s)
./cmd/mesh-controller --- PASS: TestAStopWithoutAMergeEndsTheMergesThatContainIt (0.51s)
./cmd/mesh-controller --- PASS: TestAStopWalksTheOthersAgainInTheOpenBatch (0.57s)
./cmd/mesh-controller --- PASS: TestAStopNamesWhatStaysOnTheMachinesAndWhatStillCarriesIt (0.52s)
./cmd/mesh-controller --- PASS: TestAStopWithoutAMergeLeavesItOutOfABatch (0.47s)
./cmd/mesh-controller --- PASS: TestSavingPlansAndMovingMergesIsOneAct (0.48s)
./cmd/mesh-controller --- PASS: TestNodeShowAndStatusReadTheLastSendFromTheStore (0.53s)
./cmd/mesh-controller --- PASS: TestTheConditionsVerbReadsTheLooksTheBucketNoLongerHolds (0.56s)
./cmd/mesh-controller --- PASS: TestAWalkLetGoAndStoppedByVerbsIsSaidStartedAndEnded (0.52s)
./cmd/mesh-controller --- PASS: TestAWalkStoppedOrClosedByAPersonIsSaidEndedStopped (0.50s)
./cmd/mesh-controller --- PASS: TestAWalkRetriedByAPersonIsSaidStartedAgain (0.51s)
./cmd/mesh-controller --- PASS: TestABacklogReleasedByAPersonIsSaidStarted (0.52s)
./cmd/mesh-controller --- PASS: TestAWalkEndedByCancelOrClearIsSaidEnded (0.54s)
./cmd/mesh-controller --- PASS: TestAWalkRestartedByARebuildIsSaidStarted (0.43s)
./cmd/mesh-controller --- PASS: TestAWalkStoppedByHandAndRetriedEndsFailedWhenItFailsAgain (0.53s)
./internal/inventory --- PASS: TestTheLastSendSaysWhatTheMachineAnswered (0.28s)
./internal/inventory --- PASS: TestAReportIsTheLastSendsAnswerOnlyWhenItIsAboutThatSend (0.28s)
./internal/inventory --- PASS: TestAnIssuedValueIsSealedToEveryHolderOnceAndNotAgain (0.30s)
./internal/inventory --- PASS: TestAnIssuedValueIsSealedAgainAfterAKeyChangeOrAnAcceptance (0.29s)
./internal/inventory --- PASS: TestAConditionsLooksAreKeptBesideTheBucket (0.28s)
./internal/inventory --- PASS: TestAWalksStartIsKeptAndARefusedSaveLeavesItUnsaid (0.27s)
```
Re-review against my review 64 (at 3e10ad1), now at d4de454, together with mesh-host #90 at 74ed7cb.
What I ran.go vet ./... is clean. merge-check.sh passes against a throwaway Postgres (MESH_TEST_POSTGRES set): gofmt, vet, go test -race and the 21 listed store tests all PASS. All four checks are green at this head: merge-gate (4 of 4 compose), repo-check, delivery, and delivery-group ("composed together they pass", 13:50).
The blocking finding is fixed. S21 is now judged by time since the first failing report (failing_since). It needs no report count, and it requires the machine to be heard: known, not asleep, and within its heartbeat bound. The two sides now fit:
#90 makes the node-engine say a re-apply exactly when its result changes. A recorded failure that has not been replaced therefore means the machine is still failing, so time is the right measure.
A clean re-apply is now said. recordDoing stores it as applied, NotDoingWhatTheyWereTold stops listing the machine, and the keeper clears S21 on the next tick.
Raise and clear.
TestOneReportOlderThanAnApplyIsRaisedAndAQuietMachineIsNot covers the bound: raised at 6 min, not at 4 min, not for a silent machine and not for an unknown one.
The suppression row covers the edge: not raised at exactly 5 min, raised at 5 min plus 1 s.
TestACleanReapplyAfterAFailedOneClearsTheCondition proves the clear through a real keeper.
sameFailure compares by ids, so a failure that moves between resources keeps its Since only when the ids match. Otherwise it starts again, which is what it should do.
Note, not blocking. A laptop whose failure persists across a sleep has S21 suppressed while asleep. Once it is heard again it is raised again, straight away, because Since is old. That is one warning per wake, bounded, and it is the pattern the other machine signals follow. If the channel shows it as a new notification each wake, you could carry the condition through sleep rather than letting it lapse. That would be a separate change.
Verdict: approve at d4de454
Re-review against my review 64 (at 3e10ad1), now at d4de454, together with mesh-host #90 at 74ed7cb.
**What I ran.** `go vet ./...` is clean. `merge-check.sh` passes against a throwaway Postgres (MESH_TEST_POSTGRES set): gofmt, vet, `go test -race` and the 21 listed store tests all PASS. All four checks are green at this head: merge-gate (4 of 4 compose), repo-check, delivery, and delivery-group ("composed together they pass", 13:50).
**The blocking finding is fixed.** S21 is now judged by time since the first failing report (`failing_since`). It needs no report count, and it requires the machine to be heard: known, not asleep, and within its heartbeat bound. The two sides now fit:
- #90 makes the node-engine say a re-apply exactly when its result changes. A recorded failure that has not been replaced therefore means the machine is still failing, so time is the right measure.
- A clean re-apply is now said. `recordDoing` stores it as applied, `NotDoingWhatTheyWereTold` stops listing the machine, and the keeper clears S21 on the next tick.
**Raise and clear.**
- TestOneReportOlderThanAnApplyIsRaisedAndAQuietMachineIsNot covers the bound: raised at 6 min, not at 4 min, not for a silent machine and not for an unknown one.
- The suppression row covers the edge: not raised at exactly 5 min, raised at 5 min plus 1 s.
- TestACleanReapplyAfterAFailedOneClearsTheCondition proves the clear through a real keeper.
- `sameFailure` compares by ids, so a failure that moves between resources keeps its `Since` only when the ids match. Otherwise it starts again, which is what it should do.
**Note, not blocking.** A laptop whose failure persists across a sleep has S21 suppressed while asleep. Once it is heard again it is raised again, straight away, because `Since` is old. That is one warning per wake, bounded, and it is the pattern the other machine signals follow. If the channel shows it as a new notification each wake, you could carry the condition through sleep rather than letting it lapse. That would be a separate change.
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Replaces #89/#261; same change, history without a built binary. This PR is one commit on current main. It also ignores
cmd/mesh-controller/mesh-controller, the filego buildleaves there.What
A new row S21 in the signals table adds the condition
apply-failingon the machine. It is raised once a machine's reports have said the same failure (refused whole, or applied only in part) for longer than one apply interval (5 minutes, the node-engine's ReconcileEvery), over at least two reports. It is not raised while the machine is silent or asleep, because S1 covers that, and a clean apply clears it. It reads the failing-since and count the store already keeps (issue 065,NotDoingWhatTheyWereTold), so there is no schema change.When the node-engine names a package manager that does not answer (key
the package manager (pacman), mesh-host PR on the same branch name), the condition names the tool in the summary and the headline (" cannot install packages"). The probe's words go in the evidence (Said), so the file path stays out of the summary. Otherwise it gives the refusal or the failed resource ids, with the machine's words in the evidence. Plain words are registered as a fallback.Why
Issue 233, question 3, decided 2026-10-11: a machine whose declarations are refused, or applied only in part, for longer than one apply interval raises a condition by itself (issue 187's family). On 2026-10-04 a workstation refused every declaration for 90 minutes, and the only sign was a line in
status.How it was tested
apply_failing_test.go: a machine whose package manager has not answered for 90 minutes is raised. The test checks the key, that the summary names pacman and the count and carries no path, that the evidence carries the probe's words, and the plain words. A whole refusal is also raised. A single report and a silent machine are not raised.sh merge-check.sh: green on the change (the store tests were skipped locally because no MESH_TEST_POSTGRES is set). Targeted tests re-run on current main.Fixes issue 233
Delivery
novox/mesh-controller@d4de4543ccbf— delivered since 2026-10-12T14:16:30ZDelivery plan — builds mesh-controller → novox; no bus step
Group
fix/233-a-failed-probe-fails-only-what-needs-the-tool-2, in order: novox/mesh-host@74ed7cbe0a → novox/mesh-controller@d4de4543ccMachines
Transitions
The commit's note under
refs/notes/mesh-plankeeps every transition:git log --notes=mesh-plan.Merge check at
3e10ad15mesh/merge-gate: PASS — every machine composes with the change as it did without (4 of 4 compose) (modules: build-agent, mesh-controller, route-proxy)mesh/repo-check: PASS — its merge-check.sh passedDelivery plan — builds build-agent, mesh-controller, route-proxy → ace, g14, novox, shanks; no bus step
Run by the build seat on g14 as
build-1791809446101369103(builds --log build-1791809446101369103).Merge check at
3e10ad15mesh/merge-gate: PASS — every machine composes with the change as it did without (4 of 4 compose) (modules: build-agent, mesh-controller)mesh/repo-check: PASS — its merge-check.sh passedDelivery plan — builds build-agent, mesh-controller → ace, g14, novox, shanks; no bus step
Run by the build seat on g14 as
build-1791810347306736912(builds --log build-1791810347306736912).Verdict: request changes at
3e10ad1Independent review of the issue 233 delivery group (this and mesh-host #90), at
3e10ad1.What I ran.
go vet ./...is clean.merge-check.shpasses against a throwaway Postgres (MESH_TEST_POSTGRES set): gofmt, vet,go test -raceand the listed store tests all PASS. All four checks are green at this head: merge-gate, repo-check, delivery and delivery-group.What holds.
machine.<m>.apply-failing, one per machine, so it does not flood.the package manager (key is present. The path stays out of the summary and the probe's words go into the evidence.Blocking: S21's inputs do not arrive as it assumes, so it neither counts nor clears in the live flow.
Timesadvances only when a report arrives. The node-engine's reconcile is said unasked only whenworthSaying && adoption fingerprint differs, and that fingerprint carries neither Refused nor Failed (mesh-host cmd/mesh-host/main.go, adoptionFingerprint and reconcileKept). Two results follow:"Cleared by a clean apply" therefore holds only for a clean delivery. The fix needs mesh-host #90 to say a reconcile whose outcome changed, and either:
Times ≥ 2works as written; orA test should prove the clear on a reconcile.
Merge check at
d4de4543mesh/merge-gate: PASS — every machine composes with the change as it did without (4 of 4 compose) (modules: mesh-controller)mesh/repo-check: PASS — its merge-check.sh passedDelivery plan — builds mesh-controller → novox; no bus step
Run by the build seat on novox as
build-1791812356379633013(builds --log build-1791812356379633013).Verdict: approve at
d4de454Re-review against my review 64 (at
3e10ad1), now atd4de454, together with mesh-host #90 at 74ed7cb.What I ran.
go vet ./...is clean.merge-check.shpasses against a throwaway Postgres (MESH_TEST_POSTGRES set): gofmt, vet,go test -raceand the 21 listed store tests all PASS. All four checks are green at this head: merge-gate (4 of 4 compose), repo-check, delivery, and delivery-group ("composed together they pass", 13:50).The blocking finding is fixed. S21 is now judged by time since the first failing report (
failing_since). It needs no report count, and it requires the machine to be heard: known, not asleep, and within its heartbeat bound. The two sides now fit:recordDoingstores it as applied,NotDoingWhatTheyWereToldstops listing the machine, and the keeper clears S21 on the next tick.Raise and clear.
sameFailurecompares by ids, so a failure that moves between resources keeps itsSinceonly when the ids match. Otherwise it starts again, which is what it should do.Note, not blocking. A laptop whose failure persists across a sleep has S21 suppressed while asleep. Once it is heard again it is raised again, straight away, because
Sinceis old. That is one warning per wake, bounded, and it is the pattern the other machine signals follow. If the channel shows it as a new notification each wake, you could carry the condition through sleep rather than letting it lapse. That would be a separate change.