jcoffey-dev
8725d8c11b
The public shape now works end to end: real ports, Caddy in front, and both programs that need certificates getting them from the same CA -- Caddy for the front ends over TLS-ALPN-01, the mail server for its own names over HTTP-01, which Caddy forwards on port 80. Proved in the lab against Pebble, with a DNS stub answering every name with the machine's own address, so no public name or public CA is involved: twenty checks, ending with IMAPS and submissions presenting a certificate for the mail host that verifies against the CA, and the webmail sending sign-in to the server as the first-party client the server registered. Two things the test found, both of which would have shipped: - The proxy fronted four of the server's five names. The server puts ua-auto-config in its own certificate too, so its challenge was never forwarded, one name failed, and the whole order failed with it -- leaving the mail ports on a self-signed certificate while everything else looked healthy. The list now matches what the server asks for. - Nothing waited for the certificate. An order that fails is not retried on its own and a restart does not start a new one, so the install declared itself finished over a self-signed certificate. It now waits, asks again every 45 seconds, and reports the issuer -- or says plainly that the server will keep trying once the domain resolves here, which is the ordinary case on a first install. --acme-directory and --acme-ca-root are what let a private CA be used: the root is added to the server image's own bundle and given to Caddy, because neither sees the other's trust store.
3.5 KiB
inbuxa-installer
One program that installs the inbuxa suite on the machine you run it on: the mail server, the administration console and the webmail, each as a container or on the host, in any mixture.
inbuxa survey what this machine is, as the installer sees it
inbuxa install --dry-run … what would happen, before any of it does
inbuxa install … do it
It only ever installs here. A second machine runs it too, and inbuxa join
points that machine at a server already running elsewhere.
What works today
This is early. What is built:
-
survey-- the machine's facts: distribution, init, whether Docker is usable by this user, Node's version, which of the suite's ports are free and what holds the ones that are not, memory, disk, and whether an install is already recorded here. It changes nothing. -
install --dry-run-- the whole plan: every file, unit, container, port, DNS record and credential, and a refusal with a reason when the machine cannot carry out what was asked. -
deps-- what a shape needs that this machine has not got, and, with--install, the doing of it: the Docker daemon from the distribution's own archive, the Compose plugin and Node from their official builds, both pinned by version and checked against a checksum in the source before anything is put in place.install --install-depsdoes the same as part of a run. A missing dependency is an offer, not a refusal. -
install --yes-- carries the plan out, for container shapes: writes the deployment, fetches the images, brings the mail server up in bootstrap mode with a credential that exists only for that step, completes bootstrap, brings the rest up without it, exempts the front ends from the auto-ban, creates the first mailbox, writescredentials.txtanddns.zone, and then checks that all three answer.
Not built yet: host installs, the terminal interface, join, status,
upgrade, uninstall. The design is in the inbuxa specification (§6.1 and
the installer draft); the phases are there too.
inbuxa install --local --domain example.test --install-deps --yes
is the shortest thing that works today: the whole suite on loopback, with no
DNS and no certificates, on a machine that starts with nothing. Without
--local it takes the real ports, puts Caddy in front and obtains
certificates -- which e2e/cases/install-public.sh proves against a private
CA, with no internet and no public name involved.
Building and testing
go build ./cmd/inbuxa
The installer writes units, creates users and takes ports 25 and 443, so it is tested on a throwaway virtual machine rather than on anybody's desk:
e2e/vm/up.sh a Debian 13 machine, in qemu, as you
e2e/vm/run.sh e2e/cases/survey.sh what it says about a machine
e2e/vm/run.sh e2e/cases/deps.sh the offer, and taking it
e2e/vm/run.sh e2e/cases/install-local.sh a whole suite, and signing in to it
e2e/vm/run.sh e2e/cases/install-public.sh the same with real ports and certificates
e2e/vm/down.sh remove it
Each case starts from a copy of the machine taken when it was new, so a run
is free to break it and a failure is the installer's rather than the last
run's leftovers. e2e/vm/up.sh needs qemu, KVM and xorriso; nothing needs
root on your machine.
License
AGPL-3.0-or-later. Some of this began as ihasmail-oneshot, which is ours and under the same license.