1.7 finished: minting, the file delivered, and a test flake I caused

**First, a correction: the previous commit went in on a false check.** Its message
says the suite passed; it did not. The check piped `go test` through a filter that
swallowed the failures and then printed "green" regardless. Two tests were failing
when 4de10e3 landed.

What was failing was my own doing. Purging the streams instead of deleting them
(4de10e3) left the *consumers* behind, because deleting a stream takes its
consumers with it and purging does not. A durable push consumer surviving between
tests keeps pushing to a delivery subject the previous test's subscription has gone
from: the messages count as delivered, go nowhere, and the next test waits out its
timeout for an announcement the server believes it already sent. Consumers are now
removed with the purge. Five consecutive clean runs.

`-p 1` stays, because two packages asserting and deleting the same fixed-name
objects on one bus is a real race — but its comment said the cause I had guessed
and not the one I found, so it now says the right thing.

**And delivery was not finished when I said it was.** Nothing filled
`Rendering.BusUsers`, so the composed file would never have reached a node.
`composeBusUsers` closes it: composed per push for the machine holding
`mesh-broker`, never kept, because the list is a function of the mesh's records and
a stored copy could disagree with them while both looked consistent. A user with no
credential is left out and named rather than written as a user without a password —
an ordinary situation with an obvious remedy — but a file with no users at all is
refused, because that bus would refuse every connection in the mesh.

**Minting, on both halves.** A node at enrolment and a module at `module issue`.
Three things differ from a management call and each is the point of the move: the
credential is minted into the mesh's records and becomes usable at the next
composition, so no server need be reachable; the password travels beside the address
rather than inside it, because a credential embedded in a URL leaks into every log
line that prints a connection; and a module's durable consumer is derived from what
it declared rather than named, so it cannot ask for delivery of something it did not
say it consumes.

A node reconnecting may be refused until that composition reaches the machine
running the bus. That is what the host's reconnect backoff is for and it is
survivable by design; waiting for the push would hold an enrolment open for as long
as a declaration takes to apply.

Tested that the switch is a switch: a node enrolling on one bus comes away with a
credential for that bus and none for the other, because one that held both could be
half-moved and nothing would say which half.
This commit is contained in:
2026-09-27 03:19:41 +02:00
parent 4de10e32e3
commit eb72ec36ba
8 changed files with 340 additions and 23 deletions
+71
View File
@@ -585,12 +585,19 @@ func renderingFor(ctx context.Context, open *stores, node string,
if err != nil {
return catalogue.Rendering{}, inventory.Node{}, err
}
// The bus's user list, for the machine that runs the bus. Composed per push rather than kept,
// because it is a function of the mesh's records and a kept copy could disagree with them.
busUsers, err := composeBusUsers(ctx, inv, plan.Modules)
if err != nil {
return catalogue.Rendering{}, inventory.Node{}, err
}
return catalogue.Rendering{
Settings: settings, Generators: gens, Grants: grants, Needed: needed, Ports: ports,
Certificate: certificate, Authority: authority, Mesh: private, Names: names,
Machines: machines,
Suffix: overlay.Suffix(), Foundation: foundation, Kept: kept, Adopted: record.Adopted,
Given: given, Taken: taken, Seats: seats, ArtifactStore: artifactStore, Built: built,
BusUsers: busUsers,
}, record, nil
}
@@ -1135,3 +1142,67 @@ func portsOn(
}
return out, nil
}
// composeBusUsers is the bus's user list, for a push to the machine that runs the bus.
//
// Empty for every other machine, and for every machine while the mesh is on the bus it runs on
// today — where accounts are a management call and there is no file to write.
//
// **Composed on each push, never kept.** The list is a function of the mesh's records (who exists,
// what runs where, what each declares), and a stored copy would be a second account of who may reach
// the bus, able to disagree with the records while both looked internally consistent (ADR 0043).
//
// A user the mesh has never minted a password for is **left out and said**, not written as a user
// without one — the composer refuses that, because a user with no password is a user anybody is. That
// is an ordinary situation with an obvious remedy (`module issue`, or enrolling), so the push carries
// the rest rather than failing: a bus that is missing one module's user is a mesh where that module
// cannot connect, and a bus with no file at all is a mesh where nothing can.
func composeBusUsers(ctx context.Context, inv *inventory.Inventory,
onThisNode []catalogue.Manifest) (string, error) {
_, onNATS, err := broker.OnNATS()
if err != nil || !onNATS {
return "", err
}
// Only for the machine holding the bus. Asked of what this push resolves to rather than of the
// seat's holder mesh-wide: the file is a resource of that module, so the question is whether it
// is here.
holdsTheBus := false
for _, m := range onThisNode {
if m.BusUsers != "" && m.ClaimsSeat("mesh-broker") {
holdsTheBus = true
}
}
if !holdsTheBus {
return "", nil
}
records, err := inv.BusRecords(ctx)
if err != nil {
return "", err
}
users, err := broker.Users(records)
if err != nil {
return "", err
}
kept, err := inv.BusUsers(ctx)
if err != nil {
return "", err
}
hashes := make(map[string]string, len(kept))
for name, u := range kept {
hashes[name] = u.PasswordHash
}
filled, missing := broker.WithPasswords(users, hashes)
if len(missing) > 0 {
fmt.Printf("the bus's user list leaves out %d user(s) the mesh has minted no credential "+
"for: %s. Each is a user that cannot connect until one is issued\n",
len(missing), strings.Join(missing, ", "))
}
if len(filled) == 0 {
return "", fmt.Errorf(
"this machine runs the bus and not one user has a credential, so the composed list " +
"would refuse every connection in the mesh")
}
return broker.ComposeAccounts(filled)
}