Commit Graph
314 Commits
Author SHA1 Message Date
jochen 5fc4052a2b Run a run-once process as its oneshot unit (novox/hq design 38 WP4c)
A step was run directly: in the host's own working directory, without its env, env files or
user. A module step moved out of its container (node bootstrap/index.js) could find neither its
code nor its words. A oneshot unit carries all four as a daemon's does, and starting it waits.
2026-10-04 00:29:03 +02:00
mesh-admin a7bf0f6e39 Merge pull request 'Make an unpacked archive exactly the archive (hq issue 220)' (#84) from fix/issue-220-a-bundle-on-disk-is-exactly-the-artifact into main 2026-10-03 22:17:57 +00:00
jochen 1c29168309 Make an unpacked archive exactly the archive (novox/hq issue 220)
Unpacked over the previous tree, a file the new archive no longer has stayed: a bundle rebuilt as
one file per entrypoint kept the old package directory. Unpack into a fresh directory and swap it
in, so a refused archive also leaves the old tree whole.
2026-10-04 00:17:50 +02:00
mesh-admin 483e02e7ef Merge pull request 'The installer's first user list lets the controller answer $SRV.STATS (hq ADR 0197)' (#83) from fix/0197-the-controller-answers-stats-too into main 2026-10-03 21:11:25 +00:00
jochen 0650df9414 The installer's first user list lets the controller answer $SRV.STATS (hq ADR 0197) 2026-10-03 22:47:41 +02:00
mesh-admin 47fb924947 Merge pull request 'The installer's first user list lets the controller answer discovery for its seat (hq ADR 0197)' (#82) from feat/0197-the-controller-announces-itself into main 2026-10-03 20:11:12 +00:00
jochen 09f84b451d The installer's first user list lets the controller answer discovery for its seat (hq ADR 0197)
The controller announces the mesh-controller seat on $SRV.PING/$SRV.INFO from genesis; the
carried account is the composed one, which the controller's test compares.
2026-10-03 22:11:01 +02:00
mesh-admin 77ae5320ad Merge pull request 'A process runs its own bundle's binary, written ./name (hq ADR 0193)' (#81) from feat/a-process-runs-its-own-binary into main 2026-10-03 19:14:04 +00:00
jochen b7ede83c8c A process runs its own bundle's binary, written ./name (hq ADR 0193)
A bundle compiled to a binary runs itself, but only the host knows where it unpacked it, and the
service manager takes no path relative to the working directory. A command written ./name is that
file in the process's own unpacked bundle, made absolute in the unit.
2026-10-03 21:13:55 +02:00
mesh-admin 195fc63f6a Merge pull request 'An unchanged process keeps its record, so the node's runtime is not re-created every other cycle (hq issue 210)' (#80) from fix/issue-210-a-process-is-unchanged-the-second-time into main 2026-10-03 11:41:34 +00:00
jochen 482d20737f An unchanged process keeps its record, so the node's runtime is not re-created every other cycle (hq issue 210)
The process applier's unchanged path returned an outcome that said nothing about what was
written; the loop recorded it like any other, erasing the digest. The next cycle found no
record and re-created the daemon, the one after found a record again, and so on: the
node's runtime restarted every ten minutes on every machine since it arrived. The outcome
now carries the digest forward, as a file's does. The test applies one process three
times and asserts the record survives an unchanged apply and no restart is asked.
2026-10-03 13:41:14 +02:00
mesh-admin 09614d94c8 Merge pull request 'A failed install says whether the package database is stale, how old it is, and what fixes it (hq issue 205)' (#79) from fix/issue-205 into main 2026-10-03 09:44:43 +00:00
jochen c9b963f8b9 A failed install says whether the package database is stale, how old it is, and what fixes it (hq issue 205)
pacman writes its errors to stderr, which the runner folds into the error rather than the output;
the stale-index classifier read the output alone and never saw a single 'failed retrieving file',
so a ten-week-old database on the control node reported as a wall of 404s from the mirrors. The
classifier now reads everything pacman said, knows a failed signature as the same staleness, tells a
mirror outage with a fresh database apart from it, and names the database's date and age beside the
remedy: a full upgrade by the operator, never a one-package sync, which on this distribution is a
partial upgrade. Whose job keeping the database current is stays issue 205's question.
2026-10-03 03:54:17 +02:00
mesh-admin 5a6e51f975 Merge pull request 'The first user list names the build role as node-build-agent (hq ADR 0190)' (#78) from fix/the-build-role-is-node-build-agent into main 2026-10-03 00:53:44 +00:00
jochen 3c96683d98 The first user list carries both build seats during the handover (hq ADR 0190)
Until the retired mesh-build-machine row is dropped the controller asks and hears both roles, so
the genesis list grants both; the controller's own test of this list says so.
2026-10-03 02:53:12 +02:00
jochen 8c4e33c165 The first user list names the build role as node-build-agent (hq ADR 0190)
The controller composes the build seat's subjects as node-build-agent's now; the installer's
genesis list must say the same, or the controller's own test of that list disagrees with it.
2026-10-02 22:37:08 +02:00
mesh-admin 2a332b0e6e Merge pull request 'The mesh's own ban chain is a ban wherever it hangs (hq ADR 0186)' (#76) from fix/a-ban-list-never-holds-a-neighbour into main 2026-10-02 16:43:09 +00:00
jschoubben b94e0f9a77 The mesh's own ban chain is a ban wherever it hangs; the front end's record is cited as 0180 (hq ADR 0186)
The legacy reader required every path into a chain of refusals to come from a built-in whose policy
accepts. The home server's ban chain hangs off the container runtime's user chain, whose forward
policy the runtime set to DROP, so the machine reported the mesh's own intrusion prevention as a
rule set the mesh did not write. A chain is a ban when every refusal names its sources and the
chain accepts nothing — the rule the nftables side already used. A chain that accepts anything is
still not a ban. Fixture captured from the machine. The citations for the uninstalled front end move
to ADR 0180, which another session's renumber had left pointing at an unrelated record.
2026-10-02 18:42:16 +02:00
mesh-admin b840ce0a77 Merge pull request 'A service the mesh asked to run is still running a moment later (hq ADR 0184)' (#75) from fix/a-service-asked-to-run-is-still-running into main 2026-10-02 16:25:25 +00:00
jschoubben 286865dfa7 A service the mesh asked to run is still running a moment later (hq ADR 0184)
The read-back raced the failure: a service manager returns when it has started the process, and a
daemon that refuses its configuration exits a fraction of a second later, so one look saw it alive.
fail2ban took 221ms on the control node and the apply reported "restarted" onto a dead daemon while
both public machines kept no bans at all. The host looks again, after that moment. A unit still
starting is accepted at both looks; a service asked to stop is not waited on.
2026-10-02 18:16:24 +02:00
mesh-admin ca7c4a5915 Merge pull request 'A container may log to the journal (hq ADR 0179)' (#73) from feat/the-intrusion-seat-serves-its-verbs into main 2026-10-02 15:04:02 +00:00
jschoubben b30d9c5b5a A container may log to the journal (hq ADR 0179)
A jail reads a log; a container's output went to a file of the runtime's own under a path that
changes on recreate, so no jail could read a container's service. logging: journald runs the
container with the journal as its driver, named in the spec so moving it recreates it; any other
place is refused.
2026-10-02 17:02:49 +02:00
mesh-admin 5b73e04192 Merge pull request 'A package may be declared absent, and an uninstalled front end is retired for good (hq ADR 0175)' (#71) from feat/the-found-front-end-is-uninstalled into main 2026-10-02 14:29:17 +00:00
jschoubben f57386cdea A package may be declared absent, and an uninstalled front end is retired for good (hq ADR 0175)
absent: true on a package has the host remove it through the machine's own
package manager when it is installed and leave alone a machine that never had
it; read back either way. Undeclaring a package still removes nothing. A
found firewall whose command is gone is recorded as removed, said once, and
asked nothing of.
2026-10-02 16:28:27 +02:00
mesh-admin f92dd3e286 Merge pull request 'Cite hq ADR 0170, not 0169: the firewall seat's record was renumbered' (#70) from fix/adr-0170-cited into main 2026-10-02 12:53:09 +00:00
jschoubben cdbe3ab0a4 Cite hq ADR 0170, not 0169: the firewall seat's record was renumbered after a collision on hq main 2026-10-02 14:52:20 +02:00
mesh-admin a8f1cdb445 Merge pull request 'A machine reports whether its virtualisation daemon runs (hq ADR 0172)' (#69) from jschoubben/the-lab-is-a-module into main 2026-10-02 12:47:55 +00:00
jschoubben ecb3003ba4 A machine reports whether its virtualisation daemon runs
The lab module needs the virtualisation daemon (novox/hq ADR 0172); the
capability is detected by asking the daemon about itself, not by finding
a client on disk.
2026-10-02 14:46:18 +02:00
mesh-admin d3861f82d4 Merge pull request 'A container may declare the capabilities it is granted (hq ADR 0169)' (#68) from feat/the-firewall-seat-serves-its-verbs into main 2026-10-02 11:28:34 +00:00
jschoubben b6dbe0a7b9 A container may declare the capabilities it is granted (hq ADR 0169)
Exactly the names declared reach the runtime, named in the spec so a change
recreates the container; a name that is not a capability's is refused and a
privileged container stays undeclarable. For a seat holder whose runtime
changes the machine's packet filter.
2026-10-02 13:27:34 +02:00
mesh-admin 07bdad9e94 Merge pull request 'The host says what filters the machine, with owners, and keeps the found firewall retired on every converged apply (hq ADR 0168)' (#67) from feat/one-thing-filters-a-converged-machine into main 2026-10-02 09:58:58 +00:00
jschoubben 627ac97d4f The host says what filters the machine, with owners, and keeps the found firewall retired on every converged apply (hq ADR 0168)
Every table and chain that refuses traffic is reported with whose it is:
the mesh's, the found firewall's, the container runtime's own, a ban, or
other — the runtime's user chain is other, which is where both predecessors
kept their rules, in the legacy filter on one machine and invisible to the
mesh. Adoption's threshold does not move; a converged machine's report
grows by its filters and its found firewall's state.

Convergence is a state the host keeps: a found firewall enabled again is
retired again and said; a reconcile that finds it inactive records that it
was found so, never that the mesh did it; a step skipped after a failed
apply is said. A retirement the mesh began and did not finish is finished.

Fixtures are rulesets captured from three machines of the first mesh.
2026-10-02 11:58:16 +02:00
mesh-admin ee2359f29f Merge pull request 'Every physical link faces outside, up or down (hq issue 197)' (#66) from jschoubben/every-physical-link-faces-outside into main 2026-10-02 09:54:06 +00:00
jschoubben cbdbf6b7d3 Every physical link faces outside, up or down
The filter accepts what does not arrive on a link the machine names as
outward, and the host named only links carrying a default route. An
unplugged wired port was left unfiltered for whenever it was plugged in
(novox/hq issue 197). A link backed by a physical device is now named
whether or not it is up.
2026-10-02 11:53:53 +02:00
mesh-admin e12ca3f4dd Merge pull request 'A former target of a kind the host cannot remove is left in place and said, never fatal (hq issue 194)' (#65) from fix/a-former-target-without-a-removal-is-left-and-said into main 2026-10-02 07:36:57 +00:00
jschoubben c47aa5d9b3 A former target of a kind the host cannot remove is left in place and said, never fatal (hq issue 194)
The host delivers its own successor as an archive whose target is a new
directory each version, and since mesh-host 63 the record keeps a resource's
former target for the next apply to remove. An archive has no removal (issue
162), so the first host that replaced itself under that rule refused its own
former version at the first step of every apply, and all four machines applied
nothing from then on. A former target nobody dropped is forgotten and said;
an archive the declaration dropped still refuses.
2026-10-02 02:43:47 +02:00
mesh-admin 3b9692b3cb Merge pull request 'A taken container keeps a found network, a left-out module is kept, and genesis raises the forge as its module declares (hq ADR 0163)' (#64) from feat/a-take-is-a-comparison-the-rest into main 2026-10-02 00:28:17 +00:00
jschoubben fb9c9c3ee8 A taken container keeps a found network, a left-out module is kept, and genesis raises the forge as its module declares (hq ADR 0163)
A container may name networks it also joins once created, for the per-machine
setting that keeps a found network while a neighbour still resolves it there:
joined after the run, part of the spec, refused when it cannot be joined.

A declaration may say which modules the mesh left out because a stored setting
cannot compose with its definition. Absence used to read as removal; a left-out
module's records are kept and said, and its holds are not released.

Genesis raises the bootstrap forge under the gitea module's container name, with
its image digest and its data directory mounted at /data, so the module holds it
by the found rule instead of raising a second forge beside it (issue 090). The
network is the one difference left for a take to say. Before this the forge had
no volume: its repositories were the container's, lost with it.
2026-10-01 23:45:05 +02:00
mesh-admin d53e626366 Merge pull request 'A take is a comparison: the host's facts, former targets, and strays (hq ADR 0163)' (#63) from feat/a-take-is-a-comparison-the-hosts-facts into main 2026-10-01 19:25:56 +00:00
jschoubben 83b3d20e68 A take is a comparison: the host's facts, former targets, and strays (hq ADR 0163)
Every held thing carries what a take compares: for a found container its image and the image's date,
the networks it is on and the other containers on each, its mounts and published ports, beside the
declared image (and its date once pulled), ports and volumes, with the downgrade decided when both
dates are known; for a found file whether the declared content differs and how, as lines lost and
lines new. A resource whose target moved keeps the former target on record as an orphan, so the next
apply removes the container or file the host wrote under the old name (issue 097). Every apply reports
the strays: containers the mesh neither wrote nor holds.
2026-10-01 21:24:37 +02:00
mesh-admin e030aa2387 Merge pull request 'The first user list lets the controller publish assignments and hear its seat's tools (hq #251)' (#62) from fix/first-user-list-matches-the-controller into main 2026-10-01 15:29:55 +00:00
jschoubben c670bef4e1 The first user list lets the controller publish assignments and hear its seat's tools
The controller's own composition (mesh-controller internal/broker, 2026-10-01)
publishes memberships into the assignments stream after each push and
subscribes to its seat's tool subjects; the list the installer carries did
not say so, and a controller on it is refused on the first thing it tries:
"Permissions Violation for Publish to mesh.assignment.novox.builder" (hq #251),
which is why every build asked through the console was lost. The controller's
test that compares the two (TestTheInstallersFirstUserListIsWhatTheControllerWouldCompose,
run with this checkout beside it) named exactly these two subjects, and passes.
2026-10-01 17:29:48 +02:00
mesh-admin 45529bfec2 Merge pull request 'The profile names the network manager that is running, and travels in every report (hq ADR 0161)' (#61) from feat/the-profile-names-the-uplink-and-travels-in-the-report into main 2026-10-01 14:02:16 +00:00
jschoubben b1e9ccff6d The profile names the network manager that is running, and travels in every report (hq ADR 0161)
One capability per manager — uplink-networkmanager, uplink-systemd-networkd, uplink-dhcpcd — from
systemctl is-active, so the uplink seat's holder for a manager this machine does not run is refused
the way any missing capability is, naming it (issue 138). The apply that reports detects the profile
again and sends it, the same shape enrolment sends, so a machine that switched managers reaches the
mesh at its next push.
2026-10-01 15:58:31 +02:00
mesh-admin cbcf0bcc93 Merge pull request 'A node can join the bus the mesh runs on' (#51) from fix/a-node-can-join-the-bus-the-mesh-runs-on into main 2026-10-01 11:18:02 +00:00
jschoubben 6910f07d75 Merge pull request 'Say what a container's host entries now are' (#60) from feat/148-names-are-resolved-not-copied into main 2026-09-30 12:38:15 +00:00
jschoubben 88037b33b7 Say what a container's host entries now are
novox/hq ADR 0148: the mesh's names are no longer among them, only what
the module declared. Comment only; the digest is unchanged.
2026-09-30 14:38:11 +02:00
jschoubben 422ad516d5 Merge pull request 'A declaration carries its order, and a host refuses an older one' (#59) from feat/107-a-declaration-carries-its-order into main 2026-09-30 12:03:10 +00:00
jschoubben 8431ecfb48 A declaration carries its order, and a host refuses an older one
novox/hq 04-ISSUES/107. A declaration's only identity was the digest of
its bytes: a host could say "not the last" and could not say "older". On
the link, nothing refused an older one at all, and the drain picked the
last to arrive — wrong exactly when it mattered, a backlog drained out of
order or a broker that split a burst.

A declaration may now carry a sequence, one higher per send. A host
refuses one lower than what it kept, whole, and says why. The drain keeps
the highest sequence in a batch rather than the last to arrive.

Only when both sides claim an order. Absent reads as zero — "no ordering
claimed", not "first" — so a controller that sends none is still
understood and a host that kept one before it understood them compares
nothing. That is what lets hosts go first and the controller follow, which
is the order 087 says a new field needs.
2026-09-30 14:03:03 +02:00
jschoubben f107d68b2d Merge pull request 'A delivered host knows it is the delivered one' (#58) from fix/163-a-delivered-host-knows-it-is-the-delivered-one into main 2026-09-30 11:46:58 +00:00