在 Kubernetes 容器中 KILL -9 用于构建时反应应用程序

Question

I have a react app that I want to deploy to kubernetes using a docker file and host it using nginx. My Dockerfile looks like this

FROM node:14-alpine as build

# Install git
RUN apk update && apk upgrade && \
    apk add --no-cache git

WORKDIR /src
COPY package*.json ./

RUN npm config set unsafe-perm true
RUN npm install
COPY . .
RUN npm run build

My kuberetes configuration after running kustomization looks like this

apiVersion: v1
data:
  settings.ts.ctmpl: |
    export const LIGHTHOUSE_API_HOST = 'stageBackendURL';
kind: ConfigMap
metadata:
  name: lighthouse-web-configmap-6968hk6f5g
---
apiVersion: v1
kind: Service
metadata:
  labels:
    app: lighthouse-web
    manifest_type: svc
  name: lighthouse-web
spec:
  ports:
  - port: 80
    protocol: TCP
    targetPort: 3001
  selector:
    app: lighthouse-web
  type: LoadBalancer
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: lighthouse-web
spec:
  minReadySeconds: 10
  progressDeadlineSeconds: 600
  selector:
    matchLabels:
      app: lighthouse-web
  template:
    metadata:
      labels:
        app: lighthouse-web
      name: lighthouse-web
    spec:
      containers:
      - args:
        - npm
        - run
        - serve
        image: <url to image>
        imagePullPolicy: Always
        name: lighthouse-web
        ports:
        - containerPort: 3001
          name: http
        volumeMounts:
        - mountPath: /src/src/settings.ts
          name: lighthouse-web-config
          subPath: settings.ts
        - mountPath: /src/build
          name: lighthouse-web-static
      - image: nginx:1.14
        imagePullPolicy: Always
        name: nginx
        ports:
        - containerPort: 80
          name: nginx
        volumeMounts:
        - mountPath: /usr/share/nginx/html
          name: lighthouse-web-static
      initContainers:
      - args:
        - -template
        - /config-templates/settings.ts.ctmpl:/config/settings.ts
        - -log-level
        - trace
        - -once
        image: <URL TO DOCKER IMAGE>
        imagePullPolicy: Always
        name: consul-template
        resources:
          limits:
            cpu: 500m
            memory: 500Mi
          requests:
            cpu: 500m
            memory: 500Mi
        volumeMounts:
        - mountPath: /config-templates
          name: lighthouse-web-configmap
        - mountPath: /config
          name: lighthouse-web-config
      volumes:
      - configMap:
          name: lighthouse-web-configmap-6968hk6f5g
        name: lighthouse-web-configmap
      - emptyDir: {}
        name: lighthouse-web-config
      - emptyDir: {}
        name: lighthouse-web-static

when i deploy this, the build directory is empty. I tried to ssh to the container and manually build it but got this error:

export NODE_OPTIONS=--max_old_space_size=4096 && react-scripts build

Creating an optimized production build...
The build failed because the process exited too early. This probably means the system ran out of memory or someone called `kill -9` on the process.
npm ERR! code ELIFECYCLE
npm ERR! errno 1
npm ERR! lighthouse@0.1.0 build: `export NODE_OPTIONS=--max_old_space_size=4096 && react-scripts build`
npm ERR! Exit status 1
npm ERR!

I had earlier kept the memory and cpu at 1gi and 100m but tried increasing to 16Gi and 8Gi but still didn't work.

Answer 1

Few notes on the above :

1: I'd use multistage Dockerfile build, where your whole build process happens in the node container but the final, complete, built application is then copied into an nginx based container like

FROM node as build
...
FROM nginx
COPY --from=build /path/to/app /usr/share/nginx/html

2: npm builds are expensive processes, doing npm build in your kube env, with reasonable runtime limits set, is very likely to run into issues like the one you mention here.

3: make sure that the app artifacts are correctly created on your workstation/build system