A new Go module, zfs, so the mesh can see ZFS on the home server, whose pool holds the operator's media library. Until now no mesh tool could say whether that pool is healthy, how full it is, when it was last scrubbed or which snapshots exist.
Do not assign it yet. The module is assigned to the home server only after it merges, and the first snapshot waits for the operator's yes.
The tools
Each tool says what it replaces in the manifest's replaces field (hq ADR 0245).
tool
does
replaces
zfs_datasets
Each dataset: mountpoint, used, available, referenced, compression ratio, sharenfs, its snapshot count and usedbysnapshots
zfs list
zfs_snapshots {dataset?}
Snapshots with their creation time, used (the space only that snapshot holds) and referenced
zfs list -t snapshot
zfs_snapshot {dataset, name, why, recursive?}
The only act. Details below
zfs snapshot
zfs_pools
Each pool's state, size, allocated, free, capacity, fragmentation, device errors, data errors and last scan
zpool list, zpool status -x
zfs_space
Pool space, then each dataset biggest first: the used breakdown, quota, refquota, reservation, refreservation, compression ratio, logical against physical size, and the savings
zfs list -o space, zpool list -v
zfs_pool_layout
Vdevs and disks (named by id where the pool was imported that way), each with its state and its read, write and checksum errors; log, cache, special and dedup devices; spares
zpool status, zpool status -v
zfs_scrub {history?}
The last scrub or resilver, or the progress of one running, the days since the last finished scrub, and a warning past 35 days. history also reads zpool history -i as root
zpool history -i
zfs_io {seconds?}
A zpool iostat -y -Hpv sample: operations and bytes per second per pool, vdev and disk
zpool iostat -v
zfs_arc
ARC size, target, hit ratio and L2ARC, from arcstats
arcstat, arc_summary
zfs_events {all?, limit?}
Recent errors and changes of state, as root
zpool events -v
zfs_properties {dataset}
Every property with its source. The value of keylocation is withheld
zfs get all
zfs_health
The health check
The act: zfs_snapshot
The tool refuses each of these before anything runs, and a test holds that a refusal runs no command:
a dataset that zfs list does not show;
a name with anything but [A-Za-z0-9._-];
a whole name longer than 255 characters;
an existing snapshot (with recursive, on any child too);
a missing why, or one longer than one line.
It then runs sudo -n zfs snapshot -- <dataset>@<name> as a fixed argument list, with -r only when recursive is true. It answers from zfs list afterwards: the snapshot made, its creation time, and any other snapshot that came or went during the call. A snapshot the list does not show is an error. The act is recorded as the event snapshot.taken, and ZFS's own pool history has it too. If the bus does not accept the event, the answer says NOT RECORDED.
The module has no destroy, rollback, set, send or any other change. TestTheBundleRunsOnlyReadingVerbsAndSnapshot holds every command call site to an argument function and each function to its verb. I checked that the test fails when a zpool clear is added.
Health (hq ADR 0240)
zfs_health runs on the watch process. It is unhealthy when:
a pool is not ONLINE (with status's reason and action);
a device has errors or is not online;
a pool reports data errors;
a pool is more than 90 % full;
no pool is imported;
ZFS is absent (the answer says to unassign the module).
These are warnings, which leave the module healthy: a pool more than 80 % full, and a scrub older than 35 days.
The controller's module check refuses a module that is judged only by its own tool. So beside the watch runs a second process, the event follower (zpool events -f, as root). It writes each ZFS error and change of state to its journal, where they survive a reboot (ZFS keeps events in memory only). The node-engine judges it by its unit.
What the module does not do
No seat. A new node seat is the controller's to define (ADR 0121). This module only has tools, like disk-load and sensors.
No package. ZFS belongs to the machine.
Nothing on a machine without ZFS. Every tool answers applicable: false with the reason.
Checked
go test -race ./... in modules/zfs passes. It uses OpenZFS 2.4 output fixtures (the home server runs zfs-utils 2.4.4). The fixtures are written in the exact form each command prints, not captured from the home server, because no tool could capture them yet. The first live calls are the real test of the parsers.
Writing the tests found one bug: a size error was overwritten by the compression-ratio parse. It is fixed.
merge-check.sh, run with a controller built from mesh-controller main as MESH_GATE: zfs: ok, 12 tools, ready (watch by its tool zfs_health every 1 min; events by unit every 30 s). Long-running resources without health: 70, the same as kept.
A new Go module, `zfs`, so the mesh can see ZFS on the home server, whose pool holds the operator's media library. Until now no mesh tool could say whether that pool is healthy, how full it is, when it was last scrubbed or which snapshots exist.
**Do not assign it yet.** The module is assigned to the home server only after it merges, and the first snapshot waits for the operator's yes.
## The tools
Each tool says what it replaces in the manifest's `replaces` field (hq ADR 0245).
| tool | does | replaces |
|---|---|---|
| `zfs_datasets` | Each dataset: mountpoint, used, available, referenced, compression ratio, sharenfs, its snapshot count and `usedbysnapshots` | `zfs list` |
| `zfs_snapshots {dataset?}` | Snapshots with their creation time, used (the space only that snapshot holds) and referenced | `zfs list -t snapshot` |
| `zfs_snapshot {dataset, name, why, recursive?}` | **The only act.** Details below | `zfs snapshot` |
| `zfs_pools` | Each pool's state, size, allocated, free, capacity, fragmentation, device errors, data errors and last scan | `zpool list`, `zpool status -x` |
| `zfs_space` | Pool space, then each dataset biggest first: the used breakdown, quota, refquota, reservation, refreservation, compression ratio, logical against physical size, and the savings | `zfs list -o space`, `zpool list -v` |
| `zfs_pool_layout` | Vdevs and disks (named by id where the pool was imported that way), each with its state and its read, write and checksum errors; log, cache, special and dedup devices; spares | `zpool status`, `zpool status -v` |
| `zfs_scrub {history?}` | The last scrub or resilver, or the progress of one running, the days since the last finished scrub, and a warning past 35 days. `history` also reads `zpool history -i` as root | `zpool history -i` |
| `zfs_io {seconds?}` | A `zpool iostat -y -Hpv` sample: operations and bytes per second per pool, vdev and disk | `zpool iostat -v` |
| `zfs_arc` | ARC size, target, hit ratio and L2ARC, from arcstats | `arcstat`, `arc_summary` |
| `zfs_events {all?, limit?}` | Recent errors and changes of state, as root | `zpool events -v` |
| `zfs_properties {dataset}` | Every property with its source. The value of `keylocation` is withheld | `zfs get all` |
| `zfs_health` | The health check | |
## The act: `zfs_snapshot`
The tool refuses each of these before anything runs, and a test holds that a refusal runs no command:
- a dataset that `zfs list` does not show;
- a name with anything but `[A-Za-z0-9._-]`;
- a whole name longer than 255 characters;
- an existing snapshot (with `recursive`, on any child too);
- a missing `why`, or one longer than one line.
It then runs `sudo -n zfs snapshot -- <dataset>@<name>` as a fixed argument list, with `-r` only when `recursive` is true. It answers from `zfs list` afterwards: the snapshot made, its creation time, and any other snapshot that came or went during the call. A snapshot the list does not show is an error. The act is recorded as the event `snapshot.taken`, and ZFS's own pool history has it too. If the bus does not accept the event, the answer says NOT RECORDED.
The module has no destroy, rollback, set, send or any other change. `TestTheBundleRunsOnlyReadingVerbsAndSnapshot` holds every command call site to an argument function and each function to its verb. I checked that the test fails when a `zpool clear` is added.
## Health (hq ADR 0240)
`zfs_health` runs on the `watch` process. It is **unhealthy** when:
- a pool is not ONLINE (with status's reason and action);
- a device has errors or is not online;
- a pool reports data errors;
- a pool is more than 90 % full;
- no pool is imported;
- ZFS is absent (the answer says to unassign the module).
These are **warnings**, which leave the module healthy: a pool more than 80 % full, and a scrub older than 35 days.
The controller's module check refuses a module that is judged only by its own tool. So beside the watch runs a second process, the event follower (`zpool events -f`, as root). It writes each ZFS error and change of state to its journal, where they survive a reboot (ZFS keeps events in memory only). The node-engine judges it by its unit.
## What the module does not do
- **No seat.** A new node seat is the controller's to define (ADR 0121). This module only has tools, like `disk-load` and `sensors`.
- **No package.** ZFS belongs to the machine.
- **Nothing on a machine without ZFS.** Every tool answers `applicable: false` with the reason.
## Checked
- `go test -race ./...` in `modules/zfs` passes. It uses OpenZFS 2.4 output fixtures (the home server runs zfs-utils 2.4.4). The fixtures are written in the exact form each command prints, not captured from the home server, because no tool could capture them yet. The first live calls are the real test of the parsers.
- Writing the tests found one bug: a size error was overwritten by the compression-ratio parse. It is fixed.
- `merge-check.sh`, run with a controller built from mesh-controller main as `MESH_GATE`: zfs: ok, 12 tools, ready (watch by its tool zfs_health every 1 min; events by unit every 30 s). Long-running resources without health: 70, the same as kept.
The home server's pool holds the operator's media library, and no mesh tool could say whether it is
healthy, how full it is, when it was last scrubbed or what snapshots exist; the only way was a login.
This module reads all of it with fixed zfs and zpool commands (list, get, status, iostat, events,
history) and has one act, zfs_snapshot, which only adds a snapshot: refused for an unknown dataset, a
name outside [A-Za-z0-9._-], an existing snapshot or no why; run as sudo -n zfs snapshot -- ds@name;
answered from zfs list afterwards, with any other snapshot that came or went meanwhile; recorded as the
event snapshot.taken. No destroy, rollback, set or send exists, and a test holds every command to its
verb.
Health (ADR 0240): zfs_health on the watch process, unhealthy for a pool not ONLINE, device or data
errors, a pool past 90 % full or no ZFS at all; past 80 % and a scrub older than 35 days are warnings.
Beside it, the event follower (zpool events -f, as root) keeps ZFS's errors in its journal past a
reboot and is judged by its unit, the check the module does not run itself. No seat (ADR 0121), no
package: ZFS is the machine's.
Deliverynovox/mesh-catalog@c86f0553846f — delivered since 2026-10-09T07:31:48Z
Delivery plan — builds new: modules/zfs, sent nowhere; no bus step
Machines
route-proxy on ace: passed — carried by plan-1791498071695879248
build-agent on ace: passed — carried by plan-1791499975016484055
Transitions
2026-10-08 22:16 (new) → proposed (announced): novox/mesh-catalog#146's head announced
2026-10-08 22:19 proposed → checked (checked): the verdict names this commit
2026-10-08 22:19 checked → ready (accepted): the gate passed or warned, and the repository's own check did not fail
2026-10-08 22:40 ready → held (hold): something holds it for a person: its group's composed check did not pass for the heads that merged, or it merged and no walk was opened for it
2026-10-09 07:31 held → delivered (done): merged, and nothing for a walk to move: builds new: modules/zfs, sent nowhere; no bus step
The commit's note under refs/notes/mesh-plan keeps every transition: git log --notes=mesh-plan.
<!-- mesh-delivery:view -->
**Delivery** `novox/mesh-catalog@c86f0553846f` — **delivered** since 2026-10-09T07:31:48Z
**Delivery plan** — builds new: modules/zfs, sent nowhere; no bus step
**Machines**
- route-proxy on ace: passed — carried by plan-1791498071695879248
- build-agent on ace: passed — carried by plan-1791499975016484055
**Transitions**
- 2026-10-08 22:16 (new) → proposed (announced): novox/mesh-catalog#146's head announced
- 2026-10-08 22:19 proposed → checked (checked): the verdict names this commit
- 2026-10-08 22:19 checked → ready (accepted): the gate passed or warned, and the repository's own check did not fail
- 2026-10-08 22:40 ready → held (hold): something holds it for a person: its group's composed check did not pass for the heads that merged, or it merged and no walk was opened for it
- 2026-10-09 07:31 held → delivered (done): merged, and nothing for a walk to move: builds new: modules/zfs, sent nowhere; no bus step
The commit's note under `refs/notes/mesh-plan` keeps every transition: `git log --notes=mesh-plan`.
mesh/merge-gate: PASS — every machine composes with the change as it did without (4 of 4 compose) (modules: new:modules/zfs)
mesh/repo-check: FAIL — its merge-check.sh failed: --- FAIL: TestASnapshotIsAnsweredFromTheListAfterwardsAndRecorded (0.00s)
Delivery plan — builds new: modules/zfs, sent nowhere; no bus step
Run by the build seat on novox as build-1791497529184828218 (builds --log build-1791497529184828218).
--- the gate: new:modules/zfs
zfs: ok, 12 tool(s), ready: watch by its tool zfs_health every 1m0s; events by unit every 30s
long-running resources without health: 0
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 2026-10-08T21:34:01Z 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 2026-10-08T21:34:01Z
modules the change moves:
- zfs (added)
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 0 module(s) in 0 tier(s)
--- its merge-check.sh (go toolchain)
ok git.novox.be/novox/mesh-catalog/checks/words 0.003s
words: 12 retired words, none in the 33655 strings of 600 files
ok git.novox.be/novox/mesh-catalog/checks/paths 0.007s
paths: no module names another's state on its machine (3 known, listed to go)
long-running resources without health: 70, as kept
testing modules/zfs
--- FAIL: TestASnapshotIsAnsweredFromTheListAfterwardsAndRecorded (0.00s)
snapshot_test.go:54: the creation time is the list's: 2026-10-09T12:00:00Z
FAIL
FAIL zfstools/cmd/zfs-tools 0.047s
FAIL
**Merge check** at `7c849722`
- `mesh/merge-gate`: **PASS** — every machine composes with the change as it did without (4 of 4 compose) (modules: new:modules/zfs)
- `mesh/repo-check`: **FAIL** — its merge-check.sh failed: --- FAIL: TestASnapshotIsAnsweredFromTheListAfterwardsAndRecorded (0.00s)
**Delivery plan** — builds new: modules/zfs, sent nowhere; no bus step
Run by the build seat on novox as `build-1791497529184828218` (`builds --log build-1791497529184828218`).
```
--- the gate: new:modules/zfs
zfs: ok, 12 tool(s), ready: watch by its tool zfs_health every 1m0s; events by unit every 30s
long-running resources without health: 0
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 2026-10-08T21:34:01Z 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 2026-10-08T21:34:01Z
modules the change moves:
- zfs (added)
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 0 module(s) in 0 tier(s)
--- its merge-check.sh (go toolchain)
ok git.novox.be/novox/mesh-catalog/checks/words 0.003s
words: 12 retired words, none in the 33655 strings of 600 files
ok git.novox.be/novox/mesh-catalog/checks/paths 0.007s
paths: no module names another's state on its machine (3 known, listed to go)
long-running resources without health: 70, as kept
testing modules/zfs
--- FAIL: TestASnapshotIsAnsweredFromTheListAfterwardsAndRecorded (0.00s)
snapshot_test.go:54: the creation time is the list's: 2026-10-09T12:00:00Z
FAIL
FAIL zfstools/cmd/zfs-tools 0.047s
FAIL
```
On the build node, in UTC, the fixture's creation time was the fake clock's own instant, so the test
could not tell the list's time from the clock's. A day and a half earlier is apart from it everywhere.
mesh/merge-gate: PASS — every machine composes with the change as it did without (4 of 4 compose) (modules: new:modules/zfs)
mesh/repo-check: PASS — its merge-check.sh passed
Delivery plan — builds new: modules/zfs, sent nowhere; no bus step
Run by the build seat on g14 as build-1791497801236621543 (builds --log build-1791497801236621543).
--- the gate: new:modules/zfs
zfs: ok, 12 tool(s), ready: watch by its tool zfs_health every 1m0s; events by unit every 30s
long-running resources without health: 0
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 2026-10-08T21:34:01Z 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 2026-10-08T21:34:01Z
modules the change moves:
- zfs (added)
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 0 module(s) in 0 tier(s)
--- its merge-check.sh (go toolchain)
ok git.novox.be/novox/mesh-catalog/checks/words 0.002s
words: 12 retired words, none in the 33655 strings of 600 files
ok git.novox.be/novox/mesh-catalog/checks/paths 0.005s
paths: no module names another's state on its machine (3 known, listed to go)
long-running resources without health: 70, as kept
testing modules/zfs
ok zfstools/cmd/zfs-tools 1.039s
**Merge check** at `c86f0553`
- `mesh/merge-gate`: **PASS** — every machine composes with the change as it did without (4 of 4 compose) (modules: new:modules/zfs)
- `mesh/repo-check`: **PASS** — its merge-check.sh passed
**Delivery plan** — builds new: modules/zfs, sent nowhere; no bus step
Run by the build seat on g14 as `build-1791497801236621543` (`builds --log build-1791497801236621543`).
```
--- the gate: new:modules/zfs
zfs: ok, 12 tool(s), ready: watch by its tool zfs_health every 1m0s; events by unit every 30s
long-running resources without health: 0
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 2026-10-08T21:34:01Z 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 2026-10-08T21:34:01Z
modules the change moves:
- zfs (added)
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 0 module(s) in 0 tier(s)
--- its merge-check.sh (go toolchain)
ok git.novox.be/novox/mesh-catalog/checks/words 0.002s
words: 12 retired words, none in the 33655 strings of 600 files
ok git.novox.be/novox/mesh-catalog/checks/paths 0.005s
paths: no module names another's state on its machine (3 known, listed to go)
long-running resources without health: 70, as kept
testing modules/zfs
ok zfstools/cmd/zfs-tools 1.039s
```
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.
A new Go module,
zfs, so the mesh can see ZFS on the home server, whose pool holds the operator's media library. Until now no mesh tool could say whether that pool is healthy, how full it is, when it was last scrubbed or which snapshots exist.Do not assign it yet. The module is assigned to the home server only after it merges, and the first snapshot waits for the operator's yes.
The tools
Each tool says what it replaces in the manifest's
replacesfield (hq ADR 0245).zfs_datasetsusedbysnapshotszfs listzfs_snapshots {dataset?}zfs list -t snapshotzfs_snapshot {dataset, name, why, recursive?}zfs snapshotzfs_poolszpool list,zpool status -xzfs_spacezfs list -o space,zpool list -vzfs_pool_layoutzpool status,zpool status -vzfs_scrub {history?}historyalso readszpool history -ias rootzpool history -izfs_io {seconds?}zpool iostat -y -Hpvsample: operations and bytes per second per pool, vdev and diskzpool iostat -vzfs_arcarcstat,arc_summaryzfs_events {all?, limit?}zpool events -vzfs_properties {dataset}keylocationis withheldzfs get allzfs_healthThe act:
zfs_snapshotThe tool refuses each of these before anything runs, and a test holds that a refusal runs no command:
zfs listdoes not show;[A-Za-z0-9._-];recursive, on any child too);why, or one longer than one line.It then runs
sudo -n zfs snapshot -- <dataset>@<name>as a fixed argument list, with-ronly whenrecursiveis true. It answers fromzfs listafterwards: the snapshot made, its creation time, and any other snapshot that came or went during the call. A snapshot the list does not show is an error. The act is recorded as the eventsnapshot.taken, and ZFS's own pool history has it too. If the bus does not accept the event, the answer says NOT RECORDED.The module has no destroy, rollback, set, send or any other change.
TestTheBundleRunsOnlyReadingVerbsAndSnapshotholds every command call site to an argument function and each function to its verb. I checked that the test fails when azpool clearis added.Health (hq ADR 0240)
zfs_healthruns on thewatchprocess. It is unhealthy when:These are warnings, which leave the module healthy: a pool more than 80 % full, and a scrub older than 35 days.
The controller's module check refuses a module that is judged only by its own tool. So beside the watch runs a second process, the event follower (
zpool events -f, as root). It writes each ZFS error and change of state to its journal, where they survive a reboot (ZFS keeps events in memory only). The node-engine judges it by its unit.What the module does not do
disk-loadandsensors.applicable: falsewith the reason.Checked
go test -race ./...inmodules/zfspasses. It uses OpenZFS 2.4 output fixtures (the home server runs zfs-utils 2.4.4). The fixtures are written in the exact form each command prints, not captured from the home server, because no tool could capture them yet. The first live calls are the real test of the parsers.merge-check.sh, run with a controller built from mesh-controller main asMESH_GATE: zfs: ok, 12 tools, ready (watch by its tool zfs_health every 1 min; events by unit every 30 s). Long-running resources without health: 70, the same as kept.Delivery
novox/mesh-catalog@c86f0553846f— delivered since 2026-10-09T07:31:48ZDelivery plan — builds new: modules/zfs, sent nowhere; no bus step
Machines
Transitions
The commit's note under
refs/notes/mesh-plankeeps every transition:git log --notes=mesh-plan.Merge check at
7c849722mesh/merge-gate: PASS — every machine composes with the change as it did without (4 of 4 compose) (modules: new:modules/zfs)mesh/repo-check: FAIL — its merge-check.sh failed: --- FAIL: TestASnapshotIsAnsweredFromTheListAfterwardsAndRecorded (0.00s)Delivery plan — builds new: modules/zfs, sent nowhere; no bus step
Run by the build seat on novox as
build-1791497529184828218(builds --log build-1791497529184828218).Merge check at
c86f0553mesh/merge-gate: PASS — every machine composes with the change as it did without (4 of 4 compose) (modules: new:modules/zfs)mesh/repo-check: PASS — its merge-check.sh passedDelivery plan — builds new: modules/zfs, sent nowhere; no bus step
Run by the build seat on g14 as
build-1791497801236621543(builds --log build-1791497801236621543).