diff --git a/platform/dynamic-runner/README.md b/platform/dynamic-runner/README.md index 1b241bf..f914a87 100644 --- a/platform/dynamic-runner/README.md +++ b/platform/dynamic-runner/README.md @@ -21,10 +21,13 @@ OpenSandbox chart 所有权边界。 ## Canary 安全边界 -- Deployment 为单副本,滚动策略固定 `maxSurge: 0`、`maxUnavailable: 1`,避免两个 - scheduler 共享同一个 Gitea runner 身份。 +- Deployment 为单副本,滚动策略固定 `maxSurge: 1`、`maxUnavailable: 0`,保证新 + facade Ready 后才终止旧实例。scheduler 必须通过 Kubernetes Lease 保持单 leader, + 不能依赖 Recreate 避免重复领取。 - controller 的 single-flight gate 只允许一个已领取 task 运行;Gitea 接受终态 `UpdateTask` 后才领取下一条。`POD_CAPACITY=1` 同时限制 worker 创建并发。 +- rollout 重叠期间只有持有 `Lease/dynamic-runner-scheduler` 的 controller 执行 + `FetchTask`;所有 Ready 实例都可通过 backend metadata 恢复 claim 并服务 facade。 - executor 镜像使用 digest;Pod 以 UID 2000 运行,SPIRE `ClusterStaticEntry` 同时绑定 具体 Pod UID 与 `unix:uid:2000`。 - facade 只有 ClusterIP,executor 通过 diff --git a/platform/dynamic-runner/deployment.yaml b/platform/dynamic-runner/deployment.yaml index 8e346f2..deda86c 100644 --- a/platform/dynamic-runner/deployment.yaml +++ b/platform/dynamic-runner/deployment.yaml @@ -8,8 +8,8 @@ spec: strategy: type: RollingUpdate rollingUpdate: - maxSurge: 0 - maxUnavailable: 1 + maxSurge: 1 + maxUnavailable: 0 selector: matchLabels: app.kubernetes.io/name: dynamic-runner-controller @@ -38,7 +38,7 @@ spec: mountPath: /trust containers: - name: controller - image: zot.ad.ddupan.top/panxiao81/gitea-dynamic-runner-controller@sha256:efbe44fdb13cf24f83f92d703c7051d3106cfc7d3bed9ee5debddd1d27db9dec + image: zot.ad.ddupan.top/panxiao81/gitea-dynamic-runner-controller@sha256:c165a0ec950fb5972acb9f4010c0a9cb1b89fb6c9f95c78b36cc0cbdf6f482b7 imagePullPolicy: IfNotPresent args: [controller] env: diff --git a/platform/dynamic-runner/rbac.yaml b/platform/dynamic-runner/rbac.yaml index 5dd3a7b..6f875e9 100644 --- a/platform/dynamic-runner/rbac.yaml +++ b/platform/dynamic-runner/rbac.yaml @@ -19,6 +19,9 @@ rules: - apiGroups: [""] resources: [pods] verbs: [create, get, list, watch, patch, delete] + - apiGroups: [coordination.k8s.io] + resources: [leases] + verbs: [create, get, list, watch, update, patch] --- apiVersion: rbac.authorization.k8s.io/v1 kind: RoleBinding