Use persistent Garage S3 in local deployments and CI

Replace default MinIO operator installation with pinned non-root Garage, generated persistent credentials and authenticated S3 tests. Keep MinIO manifests intact for existing installations, add failure diagnostics and optional Slack handling, and align Kustomize chart pins.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
This commit is contained in:
Cédric Verstraeten
2026-11-04 10:42:52 +00:00
co-authored by Copilot
parent 90a146a6c4
commit 7bccddb995
16 changed files with 323 additions and 54 deletions
+10 -13
View File
@@ -8,6 +8,7 @@ jobs:
deploy:
runs-on: ${{ matrix.os }}
strategy:
fail-fast: false
matrix:
os: [ubuntu-22.04, ubuntu-24.04]
#k3d: [v5.5.2, v5.6.0, v5.7.5]
@@ -29,8 +30,6 @@ jobs:
kubectl get no
kubectl get pods -A -o wide
kubectl get sc
- name: Checkout repository
uses: actions/checkout@v2
- name: Install OpenEBS
run: |
kubectl apply -f https://openebs.github.io/charts/openebs-operator.yaml
@@ -50,16 +49,11 @@ jobs:
kubectl wait --for=condition=available deployment/prometheus-operator -n default --timeout=300s
kubectl get crd
kubectl get po -A -o wide
- name: Install Minio
id: install-minio
- name: Install Garage S3
id: install-garage
run: |
git clone --depth 1 --branch v6.0.1 https://github.com/minio/operator.git && kubectl apply -k operator/
kubectl apply -f ./base/minio/minio-tenant-base.yaml
echo "Waiting for resources to be created..."
sleep 10
kubectl wait --for=jsonpath='{.status.readyReplicas}'=1 statefulset/myminio-pool-0 -n minio-tenant --timeout=300s
kubectl get po -A -o wide
kubectl get po -A -o wide | grep myminio-pool-0-0 | awk '{print $3}' | grep -q '2/2' && echo "myminio-pool-0-0 pod is running with status 2/2" || (echo "myminio-pool-0-0 pod is not running with status 2/2" && exit 1)
bash scripts/install-garage.sh
bash scripts/check-garage.sh
- name: Install MongoDB
id: install-mongodb
run: |
@@ -128,15 +122,18 @@ jobs:
kubectl wait --for=condition=available deployment/pipe-sprite -n kerberos-hub --timeout=300s
kubectl wait --for=condition=available deployment/pipe-throttler -n kerberos-hub --timeout=300s
kubectl get pods -A -o wide
- name: Collect deployment diagnostics
if: failure()
run: bash scripts/diagnose-deployment.sh
- name: Send Slack Notification on Success
if: success()
env:
SLACK_WEBHOOK_URL: ${{ secrets.SLACK_WEBHOOK_URL }}
run: |
curl -X POST -H 'Content-type: application/json' --data "{\"text\": \"✅ K3d deployment successful on ${{ matrix.os }} (${{ matrix.k3d }}) - triggered by ${{ github.actor }}\"}" $SLACK_WEBHOOK_URL
bash scripts/notify-slack.sh "✅ K3d deployment successful on ${{ matrix.os }} (${{ matrix.k3d }}) - triggered by ${{ github.actor }}"
- name: Send Slack Notification on Failure
if: failure()
env:
SLACK_WEBHOOK_URL: ${{ secrets.SLACK_WEBHOOK_URL }}
run: |
curl -X POST -H 'Content-type: application/json' --data "{\"text\": \"❌ K3d deployment failed on ${{ matrix.os }} (${{ matrix.k3d }}) - <${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|View logs>\"}" $SLACK_WEBHOOK_URL
bash scripts/notify-slack.sh "❌ K3d deployment failed on ${{ matrix.os }} (${{ matrix.k3d }}) - <${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|View logs>"
+8 -18
View File
@@ -27,7 +27,7 @@ jobs:
kubectl get pods -A -o wide
kubectl get sc
- name: Checkout repository
uses: actions/checkout@v2
uses: actions/checkout@v4
- name: Check promoted Hub runtime
run: |
helm repo add kerberos https://charts.kerberos.io
@@ -51,16 +51,11 @@ jobs:
kubectl wait --for=condition=available deployment/prometheus-operator -n default --timeout=300s
kubectl get crd
kubectl get po -A -o wide
- name: Install Minio
id: install-minio
- name: Install Garage S3
id: install-garage
run: |
git clone --depth 1 --branch v6.0.1 https://github.com/minio/operator.git && kubectl apply -k operator/
kubectl apply -f ./base/minio/minio-tenant-base.yaml
echo "Waiting for resources to be created..."
sleep 10
kubectl wait --for=jsonpath='{.status.readyReplicas}'=1 statefulset/myminio-pool-0 -n minio-tenant --timeout=300s
kubectl get po -A -o wide
kubectl get po -A -o wide | grep myminio-pool-0-0 | awk '{print $3}' | grep -q '2/2' && echo "myminio-pool-0-0 pod is running with status 2/2" || (echo "myminio-pool-0-0 pod is not running with status 2/2" && exit 1)
bash scripts/install-garage.sh
bash scripts/check-garage.sh
- name: Install MongoDB
id: install-mongodb
run: |
@@ -131,21 +126,16 @@ jobs:
kubectl get pods -A -o wide
- name: Collect deployment diagnostics
if: failure()
run: |
kubectl get pods,pvc -A -o wide
kubectl get events -A --sort-by=.lastTimestamp
kubectl -n minio-tenant describe pods
kubectl -n minio-operator logs -l name=minio-operator --all-containers=true --tail=100
kubectl -n minio-tenant logs myminio-pool-0-0 --all-containers=true --tail=100
run: bash scripts/diagnose-deployment.sh
- name: Send Slack Notification on Success
if: success()
env:
SLACK_WEBHOOK_URL: ${{ secrets.SLACK_WEBHOOK_URL }}
run: |
curl -X POST -H 'Content-type: application/json' --data "{\"text\": \"✅ Kind deployment successful on ${{ matrix.os }} (${{ matrix.kind }}) - triggered by ${{ github.actor }}\"}" $SLACK_WEBHOOK_URL
bash scripts/notify-slack.sh "✅ Kind deployment successful on ${{ matrix.os }} (${{ matrix.kind }}) - triggered by ${{ github.actor }}"
- name: Send Slack Notification on Failure
if: failure()
env:
SLACK_WEBHOOK_URL: ${{ secrets.SLACK_WEBHOOK_URL }}
run: |
curl -X POST -H 'Content-type: application/json' --data "{\"text\": \"❌ Kind deployment failed on ${{ matrix.os }} (${{ matrix.kind }}) - <${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|View logs>\"}" $SLACK_WEBHOOK_URL
bash scripts/notify-slack.sh "❌ Kind deployment failed on ${{ matrix.os }} (${{ matrix.kind }}) - <${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|View logs>"
+22 -4
View File
@@ -1,4 +1,4 @@
name: Deploy on kind using Kustmize
name: Validate Kustomize and Garage on kind
on:
workflow_dispatch:
@@ -28,16 +28,34 @@ jobs:
kubectl get pods -A -o wide
kubectl get sc
- name: Checkout repository
uses: actions/checkout@v2
uses: actions/checkout@v4
- name: Render base Kustomize manifests
run: kubectl kustomize base > "$RUNNER_TEMP/base.yaml"
- name: Use Kind storage provisioner
run: |
kubectl get storageclass standard -o json |
jq '{apiVersion,kind,metadata:{name:"ssd-hostpath"},provisioner,parameters,volumeBindingMode,reclaimPolicy}' |
kubectl apply -f -
- name: Install Garage and test persistent S3 access
run: |
bash scripts/install-garage.sh
bash scripts/check-garage.sh
credential_hash="$(kubectl -n garage get secret garage-credentials -o jsonpath='{.data}' | sha256sum)"
bash scripts/install-garage.sh
test "$credential_hash" = "$(kubectl -n garage get secret garage-credentials -o jsonpath='{.data}' | sha256sum)"
bash scripts/check-garage.sh --restart
- name: Collect deployment diagnostics
if: failure()
run: bash scripts/diagnose-deployment.sh
- name: Send Slack Notification on Success
if: success()
env:
SLACK_WEBHOOK_URL: ${{ secrets.SLACK_WEBHOOK_URL }}
run: |
curl -X POST -H 'Content-type: application/json' --data "{\"text\": \"✅ Kustomize deployment successful on ${{ matrix.os }} (${{ matrix.kind }}) - triggered by ${{ github.actor }}\"}" $SLACK_WEBHOOK_URL
bash scripts/notify-slack.sh "✅ Kustomize deployment successful on ${{ matrix.os }} (${{ matrix.kind }}) - triggered by ${{ github.actor }}"
- name: Send Slack Notification on Failure
if: failure()
env:
SLACK_WEBHOOK_URL: ${{ secrets.SLACK_WEBHOOK_URL }}
run: |
curl -X POST -H 'Content-type: application/json' --data "{\"text\": \"❌ Kustomize deployment failed on ${{ matrix.os }} (${{ matrix.kind }}) - <${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|View logs>\"}" $SLACK_WEBHOOK_URL
bash scripts/notify-slack.sh "❌ Kustomize deployment failed on ${{ matrix.os }} (${{ matrix.kind }}) - <${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|View logs>"
+11 -12
View File
@@ -8,6 +8,7 @@ jobs:
deploy:
runs-on: ${{ matrix.os }}
strategy:
fail-fast: false
matrix:
os: [ubuntu-22.04, ubuntu-24.04]
microk8s: [1.32/stable]
@@ -25,7 +26,7 @@ jobs:
kubectl get no
kubectl get pods -A -o wide
- name: Checkout repository
uses: actions/checkout@v2
uses: actions/checkout@v4
- name: Create storage class
id: create-storage-class
run: |
@@ -39,16 +40,11 @@ jobs:
kubectl wait --for=condition=available deployment/prometheus-operator -n default --timeout=300s
kubectl get crd
kubectl get po -A -o wide
- name: Install Minio
id: install-minio
- name: Install Garage S3
id: install-garage
run: |
git clone --depth 1 --branch v6.0.1 https://github.com/minio/operator.git && kubectl apply -k operator/
kubectl apply -f ./base/minio/minio-tenant-base.yaml
echo "Waiting for resources to be created..."
sleep 10
kubectl wait --for=jsonpath='{.status.readyReplicas}'=1 statefulset/myminio-pool-0 -n minio-tenant --timeout=300s
kubectl get po -A -o wide
kubectl get po -A -o wide | grep myminio-pool-0-0 | awk '{print $3}' | grep -q '2/2' && echo "myminio-pool-0-0 pod is running with status 2/2" || (echo "myminio-pool-0-0 pod is not running with status 2/2" && exit 1)
bash scripts/install-garage.sh
bash scripts/check-garage.sh
- name: Install MongoDB
id: install-mongodb
run: |
@@ -117,16 +113,19 @@ jobs:
kubectl wait --for=condition=available deployment/pipe-sprite -n kerberos-hub --timeout=300s
kubectl wait --for=condition=available deployment/pipe-throttler -n kerberos-hub --timeout=300s
kubectl get pods -A -o wide
- name: Collect deployment diagnostics
if: failure()
run: bash scripts/diagnose-deployment.sh
- name: Send Slack Notification on Success
if: success()
env:
SLACK_WEBHOOK_URL: ${{ secrets.SLACK_WEBHOOK_URL }}
run: |
curl -X POST -H 'Content-type: application/json' --data "{\"text\": \"✅ MicroK8s deployment successful on ${{ matrix.os }} (${{ matrix.microk8s }}) - triggered by ${{ github.actor }}\"}" $SLACK_WEBHOOK_URL
bash scripts/notify-slack.sh "✅ MicroK8s deployment successful on ${{ matrix.os }} (${{ matrix.microk8s }}) - triggered by ${{ github.actor }}"
- name: Send Slack Notification on Failure
if: failure()
env:
SLACK_WEBHOOK_URL: ${{ secrets.SLACK_WEBHOOK_URL }}
run: |
curl -X POST -H 'Content-type: application/json' --data "{\"text\": \"❌ MicroK8s deployment failed on ${{ matrix.os }} (${{ matrix.microk8s }}) - <${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|View logs>\"}" $SLACK_WEBHOOK_URL
bash scripts/notify-slack.sh "❌ MicroK8s deployment failed on ${{ matrix.os }} (${{ matrix.microk8s }}) - <${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|View logs>"
+134
View File
@@ -0,0 +1,134 @@
apiVersion: v1
kind: ConfigMap
metadata:
name: garage
data:
garage.toml: |
metadata_dir = "/var/lib/garage/meta"
data_dir = "/var/lib/garage/data"
db_engine = "sqlite"
replication_factor = 1
rpc_bind_addr = "127.0.0.1:3901"
rpc_public_addr = "127.0.0.1:3901"
rpc_secret_file = "/secrets/rpc-secret"
[s3_api]
s3_region = "garage"
api_bind_addr = "[::]:3900"
---
apiVersion: v1
kind: Service
metadata:
name: garage
spec:
type: NodePort
selector:
app: garage
ports:
- name: s3
port: 3900
targetPort: s3
nodePort: 30090
---
apiVersion: apps/v1
kind: StatefulSet
metadata:
name: garage
spec:
serviceName: garage
replicas: 1
selector:
matchLabels:
app: garage
template:
metadata:
labels:
app: garage
spec:
automountServiceAccountToken: false
securityContext:
runAsNonRoot: true
runAsUser: 1000
runAsGroup: 1000
fsGroup: 1000
fsGroupChangePolicy: OnRootMismatch
containers:
- name: garage
image: dxflrs/garage:v2.5.0@sha256:677e5b758e4425a325b811c425c026bf35a4bf42f7ddbaa773de3f4fd779a5a5
command: ["/garage"]
args: ["server", "--single-node", "--default-bucket"]
securityContext:
allowPrivilegeEscalation: false
readOnlyRootFilesystem: true
capabilities:
drop: ["ALL"]
seccompProfile:
type: RuntimeDefault
env:
- name: GARAGE_DEFAULT_ACCESS_KEY
valueFrom:
secretKeyRef:
name: garage-credentials
key: access-key
- name: GARAGE_DEFAULT_SECRET_KEY
valueFrom:
secretKeyRef:
name: garage-credentials
key: secret-key
- name: GARAGE_DEFAULT_BUCKET
value: mybucket
ports:
- name: s3
containerPort: 3900
startupProbe:
tcpSocket:
port: s3
periodSeconds: 2
failureThreshold: 60
readinessProbe:
tcpSocket:
port: s3
livenessProbe:
tcpSocket:
port: s3
resources:
requests:
cpu: 100m
memory: 128Mi
limits:
memory: 512Mi
volumeMounts:
- name: config
mountPath: /etc/garage.toml
subPath: garage.toml
readOnly: true
- name: credentials
mountPath: /secrets
readOnly: true
- name: data
mountPath: /var/lib/garage
- name: tmp
mountPath: /tmp
volumes:
- name: config
configMap:
name: garage
- name: credentials
secret:
secretName: garage-credentials
defaultMode: 0440
items:
- key: rpc-secret
path: rpc-secret
- name: tmp
emptyDir:
sizeLimit: 16Mi
volumeClaimTemplates:
- metadata:
name: data
spec:
accessModes: ["ReadWriteOnce"]
storageClassName: ssd-hostpath
resources:
requests:
storage: 10Gi
+4
View File
@@ -0,0 +1,4 @@
namespace: garage
resources:
- namespace.yaml
- garage.yaml
+4
View File
@@ -0,0 +1,4 @@
apiVersion: v1
kind: Namespace
metadata:
name: garage
+4 -4
View File
@@ -92,22 +92,22 @@ openai:
kerberosvault:
#uri: "http://vault-lb.kerberos-vault" # this needs to be a public accessible DNS name.
uri: http://vault-nodeport.kerberos-vault:80/api
provider: "minio"
provider: "garage"
accesskey: "XJoi2@bgSOvOYBy#"
secretkey: "OGGqat4lXRpL@9XBYc8FUaId@5"
# Archiving is used when creating a task. The underlying recording of the task will be copied from its
# existing provider to the below archived provider. Seperate credentials are used, as it makes possible to
# specify another retention period.
archive:
provider: "minio"
provider: "garage"
accesskey: "XJoi2@bgSOvOYBy#"
secretkey: "OGGqat4lXRpL@9XBYc8FUaId@5"
thumbnail:
provider: "minio"
provider: "garage"
accessKey: "XJoi2@bgSOvOYBy#"
secretKey: "OGGqat4lXRpL@9XBYc8FUaId@5"
sprite:
provider: "minio"
provider: "garage"
accessKey: "XJoi2@bgSOvOYBy#"
secretKey: "OGGqat4lXRpL@9XBYc8FUaId@5"
email:
+1 -1
View File
@@ -6,4 +6,4 @@ resources:
- factory
- vault
- hub
- minio
- garage
+2
View File
@@ -39,6 +39,8 @@ if [ "$command" == "apply" ]; then
# Adjust the base path reference
sed -i "s|../../base|./base|g" ./kustomization.yaml
bash scripts/install-garage.sh --prepare-only || exit 1
# Apply kustomize installation
kubectl kustomize ./ --enable-helm --load-restrictor LoadRestrictionsNone | kubectl apply -f -
+1 -1
View File
@@ -26,7 +26,7 @@ helmCharts:
releaseName: hub
namespace: kerberos-hub
repo: https://charts.kerberos.io
version: 0.84.0
version: 0.170.0
valuesFile: ../../base/hub/kerberos-hub-values.yaml
valuesInline:
mqtt:
+1 -1
View File
@@ -8,7 +8,7 @@ helmCharts:
releaseName: hub
namespace: kerberos-hub
repo: https://charts.kerberos.io
version: 0.84.0
version: 0.170.0
valuesFile: ../../base/hub/kerberos-hub-values.yaml
valuesInline:
kerberoshub:
+53
View File
@@ -0,0 +1,53 @@
#!/usr/bin/env bash
set -euo pipefail
if [[ $# -gt 1 || ( $# -eq 1 && "$1" != --restart ) ]]; then
echo "Usage: bash scripts/check-garage.sh [--restart]" >&2
exit 1
fi
temporary="$(mktemp -d)"
pid=""
cleanup() {
if [[ -n "$pid" ]]; then kill "$pid"; fi
rm -f "$temporary/upload" "$temporary/download" "$temporary/port-forward.log"
rmdir "$temporary"
}
trap cleanup EXIT
access_key="$(kubectl -n garage get secret garage-credentials -o jsonpath='{.data.access-key}' | base64 --decode)"
secret_key="$(kubectl -n garage get secret garage-credentials -o jsonpath='{.data.secret-key}' | base64 --decode)"
if [[ "${GITHUB_ACTIONS:-false}" == true ]]; then
echo "::add-mask::$access_key"
echo "::add-mask::$secret_key"
fi
forward() {
kubectl -n garage port-forward service/garage 13900:3900 > "$temporary/port-forward.log" 2>&1 &
pid=$!
for attempt in {1..30}; do
if curl --silent --output /dev/null http://127.0.0.1:13900/; then return; fi
if ! kill -0 "$pid"; then break; fi
sleep 1
done
cat "$temporary/port-forward.log" >&2
echo "Garage port-forward did not become ready." >&2
exit 1
}
forward
object="ci-smoke/$(openssl rand -hex 16)"
openssl rand -hex 128 > "$temporary/upload"
endpoint="http://127.0.0.1:13900/mybucket/$object"
curl --fail-with-body --silent --show-error --aws-sigv4 "aws:amz:garage:s3" \
--user "$access_key:$secret_key" --upload-file "$temporary/upload" "$endpoint"
if [[ $# -eq 1 ]]; then
kill "$pid"
wait "$pid" || [[ "$?" == 143 ]]
pid=""
kubectl -n garage rollout restart statefulset/garage
kubectl -n garage rollout status statefulset/garage --timeout=300s
forward
fi
curl --fail-with-body --silent --show-error --aws-sigv4 "aws:amz:garage:s3" \
--user "$access_key:$secret_key" "$endpoint" --output "$temporary/download"
cmp "$temporary/upload" "$temporary/download"
echo "Garage S3 authenticated PUT/GET passed through the Kubernetes Service (object mybucket/$object retained)."
+19
View File
@@ -0,0 +1,19 @@
#!/usr/bin/env bash
set -uo pipefail
kubectl get pods,pvc -A -o wide
kubectl get events -A --sort-by=.lastTimestamp
for namespace in garage mongodb rabbitmq vernemq kerberos-vault kerberos-agent kerberos-hub; do
if ! pods="$(kubectl -n "$namespace" get pods -o name)"; then
echo "::warning::Could not list pods in $namespace."
continue
fi
for pod in $pods; do
if ! kubectl -n "$namespace" describe "$pod"; then
echo "::warning::Could not describe $namespace/$pod."
fi
if ! kubectl -n "$namespace" logs "$pod" --all-containers=true --tail=80 --prefix; then
echo "::warning::Logs unavailable for $namespace/$pod."
fi
done
done
+33
View File
@@ -0,0 +1,33 @@
#!/usr/bin/env bash
set -euo pipefail
ROOT="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
if [[ $# -gt 1 || ( $# -eq 1 && "$1" != --prepare-only ) ]]; then
echo "Usage: bash scripts/install-garage.sh [--prepare-only]" >&2
exit 1
fi
kubectl apply -f "$ROOT/base/garage/namespace.yaml"
secret="$(kubectl -n garage get secret garage-credentials --ignore-not-found -o name)"
if [[ -z "$secret" ]]; then
pvc="$(kubectl -n garage get pvc data-garage-0 --ignore-not-found -o name)"
if [[ -n "$pvc" ]]; then
echo "Garage data already exists without its credentials. Restore garage-credentials from backup; refusing to generate replacement keys." >&2
exit 1
fi
kubectl -n garage create secret generic garage-credentials \
--from-literal="rpc-secret=$(openssl rand -hex 32)" \
--from-literal="access-key=GK$(openssl rand -hex 16)" \
--from-literal="secret-key=$(openssl rand -hex 32)"
fi
kubectl -n garage get secret garage-credentials -o json |
jq -e '.data | all(."rpc-secret", ."access-key", ."secret-key"; type == "string" and length > 0)' > /dev/null
if [[ $# -eq 1 ]]; then
echo "Garage credentials are ready; apply your Kustomize overlay next."
exit 0
fi
kubectl apply -k "$ROOT/base/garage"
kubectl -n garage rollout status statefulset/garage --timeout=300s
kubectl -n garage exec garage-0 -- /garage bucket info mybucket
echo "Garage is available at garage.garage:3900 (region garage, bucket mybucket). Credentials are in Secret garage/garage-credentials."
+16
View File
@@ -0,0 +1,16 @@
#!/usr/bin/env bash
set -euo pipefail
if [[ $# != 1 ]]; then
echo "Usage: bash scripts/notify-slack.sh MESSAGE" >&2
exit 1
fi
if [[ -z "${SLACK_WEBHOOK_URL:-}" ]]; then
echo "::warning::SLACK_WEBHOOK_URL is not configured; skipping optional Slack notification."
exit 0
fi
jq -n --arg text "$1" '{text: $text}' |
curl --fail-with-body --silent --show-error --retry 3 \
-H 'Content-Type: application/json' --data-binary @- "$SLACK_WEBHOOK_URL"