Reading ace's node composition gave four resources twice and no way to see that one revision named no provisioner
client's account. The verb now shows, for module node, the revision's provisioner account beside what runs names, its
adoption, how many identities, groups and libraries it names, and each module the composer's last pass held with why.
Claude-Session: https://claude.ai/code/session_01JG73KG7AfSdh6FP8VrBqCn
ADR 0593 §7: a check's composition is the checker's, made from each
placed node's answer, since only a node can compose with its own sealed
values. The node composes the indexer's candidate in shadow on a copy of
its shadow's inputs, the candidate's modules laid over the indexer's
with their manifest verdicts, and answers per module composed (changed
or not), refused or held, and its own composition's verdict, within the
answer bound; it applies, writes and hands nothing. With input it also
answers the composition input, for the sandbox path.
Claude-Session: https://claude.ai/code/session_01NcbgfJ9Gum47zkFKWRaovL
ADR 0593 §6: the provisioner, the assigner and the view read compositions
instead of composing, so the node's engine writes each one it makes to
its composition state: the wanted builds, every placed module at its
newest done build (so the plan exists before the want moves), each with
its provisioning plan, and a module refused alone with its refusal. A
key is written only when it differs, and kept while among the module's
five newest or named by a want. A placed module's trunk build landing
now makes a pass.
Claude-Session: https://claude.ai/code/session_01NcbgfJ9Gum47zkFKWRaovL
ADR 0593 §1 and §7: compose, its bus reader and runtime reader, takein,
validate, the overlay generator, the composer's run and mesh-compose, and
the judgement of a move leave mesh-sdk for the node-engine's internal
packages, so no other module composes and a change to them rebuilds only
the node-engine. The shapes others read (the composer's request and
verdict, the overlay's names, compositions and answers) stay in mesh-sdk.
The node's grant is now a golden file of what the grants part composes,
since the composer of the bus's user list leaves mesh-sdk too.
Claude-Session: https://claude.ai/code/session_01NcbgfJ9Gum47zkFKWRaovL
The bus module on the bus node runs from what its agent holds (its own composition, or the declaration kept before the
switch: every novox have read 2026-11-02 names its own composition or none, and nats names a build alone), while
composed-<node> was written by the controller, which ended (issue 1126). The follower now looks at the agent's record
of what runs whenever it changed, writes again only when the file it names moved, and trusts no composition signer.
Claude-Session: https://claude.ai/code/session_01NcbgfJ9Gum47zkFKWRaovL
Every node composes its own since group A (issue 1126), so the resync that handed the vault an own secret's value
while a node applied the controller's compositions never runs, the shadow's answers read no want, and nothing follows
the heard vault key any more.
Claude-Session: https://claude.ai/code/session_01NcbgfJ9Gum47zkFKWRaovL
The controller ended by its unassignment, so a node that falls back to the
compositions in its bucket follows what nobody writes (issue 1126, A7). The
node agent now always runs its own composer: without a placement key it
composes nothing and keeps what runs, a placement it cannot watch stops the
agent and is said, and a candidate taking its placement away is refused
instead of answered as a return to the controller. The start-time placement
read, the end-when-placed watchers and the returning answer go, with their
tests and the three integration tests of the controller-composed path.
Claude-Session: https://claude.ai/code/session_01NcbgfJ9Gum47zkFKWRaovL
The controller ended by its unassignment and its lease is never handed, so
the engine's lease gate kept trusting the controller's key and the bus node
kept the controller's list as a fallback nobody writes (issue 1126, A6).
The engine now trusts no controller signature, its hand-over line names the
node seat's hand-over, and the user-list follower writes only the grants
part's list: a key unreadable or deleted keeps the list loaded, and with no
key heard nothing is written. Compositions in the node's bucket name the
file's place alone. The guard no longer asks for the controller's user.
Claude-Session: https://claude.ai/code/session_01NcbgfJ9Gum47zkFKWRaovL
pacman still reads the controller's mirror because no engine wrote its key: the engine
now keeps each handed composition's package resources and, after a pass that handed
something new, writes mesh-host_packages.<node> only when the value differs, owed after
the growing bound when refused. The manifest declares the state so the grants part
creates it. The verb packages answers the key beside the mirror's key for the node and
their difference, and with write writes the key again from what was composed. H8's
integration test proves the write under the grant the grants part composes.
Needs mesh-sdk feat/node-packages (0ef7c0c) for DeclaredPackagesOf and the engine's read of the
mirror (go.mod pin left for the lead: the SDK commit is not pushed).
Claude-Session: https://claude.ai/code/session_01NcbgfJ9Gum47zkFKWRaovL
After a pass that handed a new composition, the node's composer works out its modules' data
items from the inputs it composed from and writes each module's key <node>.<module> of
mesh-host_data only when it differs, deleting the keys of modules that left. Pins mesh-sdk
feat/data-items for the contract and the engine's grant. Two tool-answer tests are merged so
the unit count stays where it was.
Claude-Session: https://claude.ai/code/session_01NcbgfJ9Gum47zkFKWRaovL
On novox (2026-10-31) postgres's repeat backup was interrupted by the node-engine's
restart: its held have named no build, so the node's own composition took postgres in
and was refused for its secret reader, and nothing ran the backup again. Issue 1289
(R4c, its delivery held).
- A held have names what runs: read from the module's have before backing-up when
the agent has no running entry, as the not-composed mark does.
- A have in any phase but running that names nothing is not run inside the node's own.
- A backup interrupted by the agent's stop is no failure: the next start backs the
module up again and applies while the want asks a backup. A real failure stays
held for the operator's answer.
Quick fix: a new backup ask by hand. Thorough, chosen: the fault fixed where it lives
in the node agent, so every data module's interrupted backup resumes by itself.
Claude-Session: https://claude.ai/code/session_01NcbgfJ9Gum47zkFKWRaovL
On 2026-10-31 ace's route-proxy said only that the vault could not be asked, and ace said nothing for five minutes
while a pass waited after the vault restarted; the retry that already existed was invisible. A refusal for a seat or
a read that did not answer now says when the node composes again ("the node composes again at HH:MM:SS (try N)"), a
retry pass says it starts, a pass that runs on says every minute what it waits on, and the node-engine's verb
compose makes a pass now and says how it went, so a person recovers without a restart.
restic's composition dropped its process restic.measure on g14 and shanks at 16:14; the engine removed it
and said it down on every look after. The judges were set only when facts were said, and the facts said
inside an apply read the module's composition from before it (the agent records the new one only after
the apply returns), so the removed resource was put back. Judge.Set itself prunes what is not declared.
Quick fix: set the judges after the restic apply. Thorough fix (chosen): one function hands every judge
(liveness, resolver, units, accounts, libraries) what is declared; it runs from the facts as before and
now also whenever a module settles (applied or removed) or is taken as running, so every removed health
entry of every module is pruned at recompose.
On 2026-10-30 ace's redis, postgres and restic (14:03), step-ca (13:55) and mesh-vault on novox (15:22)
were held on a single backup ask that met no responder while the tool runner restarted.
Quick fix: ask node-backup.now twice. Thorough fix (chosen): every node-backup ask (backed-up and now)
is asked again with the engine's backoff while it is known never to have been delivered (ErrNotAsked),
bounded at three minutes; an ask that may have been delivered (a timeout, a refusal) is never asked
twice, so no backup starts twice. Past the bound the backup fails saying how often and how long it was
asked, the module is held and its have carries those words for the assigner's ask and condition. This
was the engine's only one-shot seat ask.
delivery.NetworkPart.Targets (mesh-sdk 6eae076) is filled where the report is made, by the controller's own rule:
- the hub's node for "hub";
- each resolver's node by its address on the private network;
- none when the part points at nothing, at this machine, or at an address no other mesh machine has.
The names come from the placements the node's own composer holds (Decided.Address, Decided.Hub). A spoke's tunnel names
only the hub, so the tunnel cannot name a resolver on another spoke. A node that does not compose its own names no
target.
Test: rows in TestAHealthStatementIsSaidOnTheLinkOpenNow (hub, two resolvers, a VPN's resolver, its own, a healthy
part).
The reviewer showed how 76e214a still lets the 07:54 outage happen at the re-landed engine's first start on novox.
applied/ is empty there, so loadApplied loads nothing. The kept declaration retires, because every placed module has a
have. Nothing then carries the node's own composition. The guard misses nats, which has no want there, so its
nats.conf is guessed as nats's and removed.
- At the start, when the store has no node's own composition, the node's own runs as the kept declaration has it, and
the store keeps it now. The kept declaration retires only once the store holds `node`.
- What this engine applied runs until loadHad reads the composition a differing have names. The bucket no longer keeps
the one novox's have names, so the stored one stays.
- In the node's own apply, a resource of another module goes only when that module's want says removed
(Scope.RemovedByWant). A module the node's own composition carried has no want, so Placed never named it.
Cost: the node's own apply cannot drop one of its own dotted resources without such a want. It is refused whole and
said, never removed silently.
Tests, with no new test functions:
- A first-start case in the agent's Everything test (empty store, then a start with the declaration retired). It fails
on 76e214a.
- Two rows in the guard's table: no want refuses; a want saying removed removes.
- A case for the store in the authority test.
On novox on 2026-10-29 at 07:54 UTC the kept declaration was retired and the node's own composition was in no bucket.
The composer's first pass reads what runs before the agent reads any have (the agent waits for the composer's caught,
sent after that pass), so the node's own was composed with its folded modules running nothing, and the bus server and
four held modules were removed.
- Quick fix: read the node's own have in loadHad too. It still waits for the composer's first pass, and still finds
nothing when the bucket no longer keeps the composition the have names.
- Thorough fix, chosen: the engine keeps each composition it applied (the node's own too) in `applied/` beside its
state, 0600. The agent reads it at Run's start, before any watch, as what runs ahead of the bucket and the kept
declaration. A have naming another running composition wins. An apply keeps; a removal forgets.
Test: a case in TestEverythingIsWhatTheAgentRunsAndWhatIsKeptForTheRest (no new test function). It fails on L3's
engine 4f9227a (0 compositions at the first own composition) and passes here.
novox's mesh-vault has failed since 01:16 on 2026-10-29: its have says failed, so loadHad (fbea777) takes no
composition from it, and what runs falls back to the kept declaration, whose export (mesh-vault.kept) names
f099eaa1, an object the controller's sweep removed. The node wrote a newer export since. The composer carries that
file as it runs (mesh-sdk frombus keptOf), so what runs now names the digest and size of what is on disk, but only
where the node's record says it wrote exactly that content at that path (Applied.Wrote, Target). The apply then
finds the file in place and needs no object.
Quick fix weighed: have the controller put the object back. Thorough (chosen, in the node): what runs is what the node
wrote, never a guess from an older declaration. The rows join the existing whole-apply test; the test count stays.
The kept-declaration fallback built a composition without Data, so after a restart a data module (mongodb, mssql
on novox) was judged to move its data against its first composition and the operator was asked a false yes. Now
every such composition takes Data from the manifest (agent.Env.HoldsData, the composer's DataItems), and a module
whose manifest cannot be read is said to be unjudgeable rather than taken to hold none. The composition a have
names comes before the kept declaration in the judgement too.
loadHad takes a have only in the phases baseline takes as running (running, held, a composer's mark): one backing
up, applying, checking, failed or waiting names a composition not applied (review of 031e106). Two rows added to
the same table test; the count stays.
A held module on a node composing its own never gets a composition, so the agent never took it and the filling
judged its files by the kept declaration, older than power's three new drop-ins: refused from every engine start
on ace, g14 and shanks until some module settled (2026-10-29 00:09 UTC). The second meeting of issue 924.
Quick fix: knock the filling once the haves are read at the start. Thorough fix (chosen, the kind recurs for every
file added since the switch): what runs of such a module is the composition its have names, from the bucket or
its revisions, ahead of the kept declaration; the filling is knocked on every change of what runs, not only on a
settle. One unit test replaced by a table of four cases; the count stays.
The answer names its proposal and the revision of the node's own candidate key; a node with no key in a proposal is not asked. The watch for its own placement key is tried again until the seat's state exists, so the node switches when its key is written.
mesh/check being carried to bf667a602d21 on 58b4f8b6bc07 | from 9a8b3870bad0
A reboot left the have running, so a unit that failed to start at boot was never started again and nothing said so. Baseline now asks the machine; the judge's look calls Agent.Reconcile once per down edge, leaving held and unwanted modules alone.
mesh/check pass 24c23a1d066f on a548a61e86d7 | unit 2667 run, 590 kept, 36 skipped | integration 2676 run, 633 kept, 29 skipped | over bound: unit (sl…
frombus.Compose now holds a module refused alone (mesh-sdk 0de3006, the rule
its answer follows too): composed as what runs of it, with what depends on
how it is held held too. The composer hands every other module, keeps what it
handed of a held one, and tells the agent which modules it holds and why;
agent.NotComposedEach marks those alone and clears the rest. Only a refusal
of no single module still marks every module (NotComposed).
heldOut also holds out, of the node's own composition as it runs, the
resources of each module held alone, until a composition of its own is
handed: the node's own apply never removes a module leaving the fold with its
names unset (review of the apply, round 1; issue 1945 by a new path).
Pins mesh-sdk fix/1126-apply-culprit at 0de3006 (a group).
Met twice on 2026-10-28: shanks kept 'want.shanks.brother could not be read ... no
responders' as the own-secret step's last word, and ace kept 'the vault's own: not yet
... no responders' as its answer to seed-7, each until another input moved. Every bus
read of the composer (manifests, builds, archives, served values, the vault's own and
credentials, the have of what runs) is noted (BusFailed); an answer or a pass that met
one is made again after a growing bound (5 s to 5 min), and a pass is not handed when
the have of its sealed values did not answer. The own-secret step tells a read or an
ask that did not answer (Retry, taken again) from a key absent or the vault's word.
A new step, internal/ownresync, after each shadow answer: for each own secret masked as another value, while the node
applies the controller's compositions and runs the one its want names, it asks the vault's move-in with
replace_running: true and the running value sealed to the call's key, once per value, and says what it asked and what
the vault answered. The answer written to the node's have key carries values_differ. Pins the group's mesh-sdk head
1062b33 (go.sum is filled once that head is pushed).
A box is reused by slot (resource, field) after the composition, from what runs or the
have's slots; a have written before slots is reused within its module alone. Config
values are laid by mesh-sdk's one rule (never a default). The shadow answers the
newest candidate first and hands each answer as it lands, says when an answer starts,
every minute what it waits on, and when it ends, and fits an answer to the node's key.
Needs mesh-sdk fix/1126-seed-same (16b8d8a1).
Live 2026-10-28: the own run's last pass, failing on its run's end, marked
every module of the controller-mode agent that replaced it (ace 60, shanks 63),
and the marks dropped the build and digest because the have read failed in the
bus's reload pause. Now: the refusal hook is bound to the run's agent and ctx;
a pass whose context ended returns nothing; JetStream not enabled or no
responders are tried again after 1, 2 and 4 s before any mark; NotComposed reads
the have again while the bus does not answer and marks nothing it cannot read;
baseline takes a have a composer marked as what runs, so a mark it did not make
clears; each run's shadow lives with the run.
mesh/check checking 463626143567 on bb15aab50eda | novox/mesh-host#154 in group fix/1126-shadow-all
selfComposer took a KVHistory error for 'not placed': at any restart a timeout
or a bus not ready flipped a self-composing node to the controller's
compositions for a whole run. selfcompose.Placed chooses only on a read that
succeeds (a missing placement state is one), and tries again otherwise, said
once. The shadow's answer to a head deleting the placement reads the
controller's composition at each want in the node's bucket and compares it with
what runs; a want with none refuses, without returning.
The seed's revert (mesh-catalog #526) was refused on 'would stop running node',
while a self-composing node would have refused its placement's absence, kept
what runs and marked every have failed until a restart. endWhenUnplaced ends the
agent's run when the key goes, so it begins again on the controller's
compositions, as at a start without a key; the composer refuses nothing while
unplaced; the node's have says whose compositions it applies (Composes); the
shadow answers a deleted placement Returns, verdict same.
The node read a behind module's source manifest digest and composed every module
from the indexer's raw manifest; frombus now takes the build itself (Scope.KeptAt,
a fingerprint). The reader gives frombus the builder's key, a templates archive
fetched as every archive is (apply.Fetch) once per digest and checked against it,
and another node's served keys one by one, as the node's grant allows them.
mesh/check failed: unit: ..test: ✔ a transient empty list for four passes retires nothing; five stable passes retire, data kept (290.408814ms) | 58de6…
mesh/check failed: ace: novox/mesh-catalog#505: checks/placement/go.mod names mesh-sdk 91413de33516, which is neither on mesh-sdk's trunk nor the grou…
jschoubben