尚未偵測到更新。GitHub 發布版本後會顯示在這裡。
個人 GitHub 帳號下公開儲存庫的擁有者可以登入並驗證所有權,自行管理介紹,無需管理員核准。
登入並產生驗證檔案,在 30 分鐘內將其提交至此儲存庫的預設分支。其他分支或分叉儲存庫中的檔案無效。然後在下方檢查檔案,即可無需管理員核准直接管理專案。請保留檔案以維持管理權限。
如需登入,將前往登入頁面。登入後請回到此處,再次點擊所選的驗證按鈕。
在專案資料夾中執行 /nomorevibe verify,即可驗證所有權並自行管理介紹。
一分鐘完成驗證Tooling, scripts, and operating standards for running IBM MQ high-availability queue managers on Ubuntu Linux clusters.
翻譯準備中
取自儲存庫 README
正在載入完整 README。 以下為已儲存的摘錄。
A downloadable, signed lab that stands up IBM MQ high-availability (HA) and
disaster-recovery (DR) topologies end to end — from a bare x86 Linux host down
to running queue managers and a live message path that survives failover. You
bring a virtualization-capable host; the lab brings everything else, and stands
up a full HA/DR stack with one command.
Full documentation lives in the site under docs/site/ —
this README is the on-ramp; the site is the depth.
• What this is
• Getting Started
• Prerequisites
• Download and verify the release
• Run it
• Developing with Vergil
• Development檢視完整 README ↗A downloadable, signed lab that stands up IBM MQ high-availability (HA) and disaster-recovery (DR) topologies end to end — from a bare x86 Linux host down to running queue managers and a live message path that survives failover. You bring a virtualization-capable host; the lab brings everything else, and stands up a full HA/DR stack with one command.
Full documentation lives in the site under docs/site/ —
this README is the on-ramp; the site is the depth.
At the core is mqlab, a standalone Python orchestrator, plus its support
tree — the Ansible roles that provision everything from OS to running queue
managers, the lab/topology.yaml fleet definition, and the
release manifests. The deliverable is the whole tree, not a Python module:
you run it from the extracted, signed release.
mqlab does the opposite of most tooling — rather than hiding the mechanics, it
shows them. Every command it runs (virsh, Ansible, runmqsc) is printed
verbatim as it runs and teed to a transcript, so you can watch a step, understand
it, then reproduce it by hand. The lab is transparent evidence, not a black box.
The author develops the lab inside a personal
Vergil dev VM, but Vergil only provides that
dev environment — strip it away and what remains is exactly mqlab and its
support tree, which is what the signed release ships and what you run on your own
host. See Developing with Vergil below.
This is the orientation-altitude version of the site's Getting Started — start there for the full walkthrough.
Almost everything is fetched automatically — IBM MQ Advanced for Developers
(the no-charge edition) is pulled by the tooling, at the level pinned in
lab/mq-version (currently 10.0.1.0). You provide:
mqlab doctor reports anything missing.uv provisions its own pinned 3.14; the base OS Python is untouched).pcmk-ubuntu arm needs nothing extra. Both RHEL arms
default to RHEL 10, which needs an x86-64-v3 CPU (mqlab doctor says
whether yours is one); without it, build them on RHEL 9 instead.Each stack's default and selectable OS versions (Ubuntu 24.04/26.04, RHEL 9/10)
live in one catalog, lab/versions.yaml; the site's
Getting Started shows
how to pick one.
The lab ships as a signed tarball from the Releases page. Download the tarball and the two checksum files (substitute the version you're installing):
VERSION=vX.Y.Z
base=https://github.com/logical-minds-foundry/mq-resiliency-lab-for-linux/releases/download/$VERSION
curl -LO $base/mq-resiliency-lab-for-linux-$VERSION.tar.gz
curl -LO $base/SHA256SUMS
curl -LO $base/SHA256SUMS.asc
Verify it before you unpack. The trust root is the fingerprint below plus a key fetched out-of-band — not any key bundled in the tarball:
gpg --recv-keys 5BABD50A78EBF24D2410AE52ADF1A99B75D24E54 # or import the key from the release page over HTTPS
gpg --fingerprint 5BABD50A78EBF24D2410AE52ADF1A99B75D24E54 # cross-check it matches this README
gpg --verify SHA256SUMS.asc SHA256SUMS # the checksums are signed by that key
sha256sum -c SHA256SUMS # the tarball matches the checksums
Once both checks pass, unpack the tarball and enter the tree:
tar xzf mq-resiliency-lab-for-linux-$VERSION.tar.gz
cd mq-resiliency-lab-for-linux-$VERSION
From inside the extracted tree, build the environment, pre-flight the host, then bring up your first stack:
./scripts/setup # checks uv + Python 3.14, then runs `uv sync` to build the venv
. .venv/bin/activate # put `mqlab` (and its companion tools) on PATH
mqlab doctor # pre-flight the host: arch, KVM, libvirt, required tools
mqlab bootstrap pcmk-ubuntu # bring up a full HA/DR stack: net → vms → provision → observe
mqlab status pcmk-ubuntu # phase completion (✓/✗) for this stack
./scripts/setup sets up the environment, not the lab: it verifies uv and
a uv-managed Python 3.14 are present (installing 3.14 via uv if missing), runs
uv sync to materialize the
virtual environment, and prints the next commands. It cannot change your shell's
PATH, so activate the venv (or prefix each command with uv run). mqlab bootstrap pcmk-ubuntu
then sequences the whole bring-up — the canonical starting arm is pcmk-ubuntu
(Pacemaker/SAN on Ubuntu, subscription-free and x86-native) — and leaves you with
a running queue manager. See
Getting Started to send your first message
and watch it survive a failover.
The lab is developed inside an ephemeral Vergil VM — the author's personal, solo-practitioner dev tooling — rather than on the host directly. That is a development convenience, not part of the consumer path above: strip Vergil away and what remains is exactly the tree you unpacked. It's my own tooling; if you'd find it useful, talk to me.
Developing the lab itself (as distinct from running it) uses the Vergil
workflow. Footprint and tooling are declared as the [vm.vergil-user] profile in
vergil.toml; build and enter the box with:
vrg-vm create logical-minds-foundry/mq-resiliency-lab-for-linux --identity vergil-user
vrg-vm session logical-minds-foundry/mq-resiliency-lab-for-linux --identity vergil-user
The box is ephemeral and fully reproducible, so don't hand-customize it.
Nothing in it outside the host-mounted build/ directory is precious;
re-provision it freely to stay fresh:
vrg-vm rebuild logical-minds-foundry/mq-resiliency-lab-for-linux --identity vergil-user
To change its tooling, edit the [vm.vergil-user] profile and rebuild; never
apt install inside a live VM.
From a dev checkout, run the CLI as uv run mqlab … rather than calling the
venv's mqlab by its path: uv run puts the venv's bin/ on PATH, so the
tools mqlab calls by bare name resolve (see
docs/development/operating-the-lab-from-a-dev-session.md).
See CLAUDE.md for the workflow,
docs/site/docs/develop.md for the developer
on-ramp, docs/development/release-runbook.md
for cutting a release, and
docs/development/perf-and-staging.md
for comparing bootstrap perf reports across macOS and x86 cloud
(mqlab perf diff, the auto-detected MQLAB_ENV profiles, macOS huge pages, and
measured bootstrap times).