问题
I am trying to deploy a sample angular app to GKE. I created a sample cluster enabling cloud run and istio services in it
gcloud beta container clusters create new-cluster \
--addons=HorizontalPodAutoscaling,HttpLoadBalancing,Istio,CloudRun \
--machine-type=n1-standard-2 \
--cluster-version=latest \
--zone=us-east1-b \
--enable-stackdriver-kubernetes --enable-ip-alias \
--scopes cloud-platform --num-nodes 4 --disk-size "10" --image-type "COS"
Following is my cloudbuild.yaml file steps:
# build the container image
- name: gcr.io/cloud-builders/docker
args: [ build, -t, gcr.io/$GCLOUD_PROJECT/gcp-cloudrun-gke-angular:1.01, . ]
# push the container image to Container Registry
- name: gcr.io/cloud-builders/docker
args: [ push, gcr.io/$GCLOUD_PROJECT/gcp-cloudrun-gke-angular:1.01 ]
# Deploy container image to Cloud Run
- name: gcr.io/cloud-builders/gcloud
args: [ beta, run, deploy, feedback-ui-deploy-anthos, --image, gcr.io/$GCLOUD_PROJECT/gcp-cloudrun-gke-angular:1.01, --platform, gke, --cluster, cloudrun-angular-cluster, --cluster-location, us-central1-a ]
images:
- gcr.io/$GCLOUD_PROJECT/gcp-cloudrun-gke-angular:1.01
I have set the env var for gcloud prj. Now when I try deploying this tothe gke cluster created above, I always end up with the revision unavailable error:
Deploying new service... Configuration "service-1" does not have any ready Revision.
- Creating Revision...
X Routing traffic... Configuration "service-1" does not have any ready Revision.
This is the command I used to deploy to cloud run
gcloud beta run deploy --platform gke --cluster new-cluster --image gcr.io/$GCLOUD_PROJECT/gcp-cloudrun-gke-angular:1.01 --cluster-location us-east1-b
The other fully managed cloud run works perfectly. But when I deploy to existing gke clusters, I end up with the error. I read through the documentation and it says that the revision gets created automatically if it is a new service, not sure why that is not happening for my service
EDIT: Here is the kubectl describe output. I deleted all clusters and created a new one afresh and still ended up with the same.
So upon describing the service, this is what I get
Note: I use the default namespace. not sure if it has any bearing on this issue.
Status:
Conditions:
Last Transition Time: 2019-12-04T12:49:59Z
Message: Revision "gke-service-00001-pef" failed with message: Container failed with: nginx: [alert] could not open error log file: open() "/var/log/nginx/error.log" failed (2: No such file or directory)
2019/12/04 12:49:40 [emerg] 1#1: open() "/var/log/nginx/error.log" failed (2: No such file or directory)
.
Reason: RevisionFailed
Status: False
Type: ConfigurationsReady
Last Transition Time: 2019-12-04T12:49:59Z
Message: Configuration "gke-service" does not have any ready Revision.
Reason: RevisionMissing
Status: False
Type: Ready
Last Transition Time: 2019-12-04T12:49:59Z
Message: Configuration "gke-service" does not have any ready Revision.
Reason: RevisionMissing
Status: False
Type: RoutesReady
Latest Created Revision Name: gke-service-00001-pef
Observed Generation: 1
URL: http://gke-service.default.example.com
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Created 2m21s service-controller Created Configuration "gke-service"
Normal Created 2m21s service-controller Created Route "gke-service"
Normal Updated 20s (x5 over 2m21s) service-controller Updated Service "gke-service"
Since I am exposing the angular index.html file via nginx, this is my configurtion:
server {
listen 8080 default_server;
sendfile on;
default_type application/octet-stream;
gzip on;
gzip_http_version 1.1;
gzip_disable "MSIE [1-6]\.";
gzip_min_length 1100;
gzip_vary on;
gzip_proxied expired no-cache no-store private auth;
gzip_types text/plain text/css application/json application/javascript application/x-javascript text/xml application/xml application/xml+rss text/javascript;
gzip_comp_level 9;
root /usr/share/nginx/html;
location / {
try_files $uri $uri/ /index.html =404;
#proxy_pass: "http://localhost:8080/AdTechUIContent"
#uncomment to include naxsi rules
#include /etc/nginx/naxsi.rules
}
}
This works fine when I build the docker image in local and I able to access it. Just in case, this is my docker file
FROM node:12.13-alpine as app-ui-builder
#Now install angular cli globally
RUN npm install -g @angular/cli@8.3.14
#RUN npm config set registry https://registry.cnpmjs.org
#Install git and openssh because alpine image doenst have git and all modules in npm has the dependicies which are all uploaded in git
#so to use them we need to be able git
RUN apk add --update git openssh
RUN mkdir ./app
COPY package*.json /app/
WORKDIR ./app
COPY . .
RUN npm cache clear --force && npm i
RUN ls && $(npm bin)/ng build --prod
FROM nginx:1.17.5-alpine AS nginx-builder
RUN apk update && apk add ca-certificates && rm -rf /var/cache/apk/*
COPY app-ui-nginx.conf /etc/nginx/conf.d
RUN rm -rf /usr/share/nginx/html/*
COPY --from=app-ui-builder /app/dist/app-ui /usr/share/nginx/html
RUN ls /usr/share/nginx/html
RUN chmod -R a+r /usr/share/nginx/html
EXPOSE 8080
#
CMD ["nginx", "-g", " daemon off;"]
@AhmetB . Could you let me know why the nginx is throwing error here
EDIT: I did try deploying the app using plain Kubectl commands with a deployment and a service. It worked fine. So not sure what cloud run contract it violates for the error logging with nginx even though the file can be found
回答1:
Deploying new service... Configuration "service-1" does not have any ready Revision.
This error means that it's deployed but for some reason the pod is crashing or not scheduling. This can happen for all sorts of reasons like not enough CPU/memory on the node, image cannot be pulled from GCR, or the app is crashlooping.
Look at "kubectl logs" and "kubectl describe" output for your application. Try:
- kubectl get ksvc
- kubectl get pods
- kubectl describe ksvc NAME
- kubectl logs NAME -c user-container
回答2:
I found the issue. Looks like the log files (both error and access log files) should be created in a custom folder for cloud run to access to. Cloud run checks for those folders to be available before spining up a revision. When I used the old nginx config file, no custom folders were created. Now modified the nginx conf file and deployed it and it worked fine
Created two files nginx.conf
user nginx;
worker_processes 1;
error_log /var/logs/nginx/error.log warn;
pid /var/run/nginx.pid;
events {
worker_connections 1024;
}
http {
include /etc/nginx/mime.types;
default_type application/octet-stream;
log_format main '$remote_addr - $remote_user [$time_local] "$request" '
'$status $body_bytes_sent "$http_referer" '
'"$http_user_agent" "$http_x_forwarded_for"';
access_log /var/logs/nginx/access.log main;
sendfile on;
#tcp_nopush on;
keepalive_timeout 65;
#gzip on;
include /etc/nginx/conf.d/*.conf;
}
default.conf
server {
listen 8080;
server_name localhost;
location / {
root /usr/share/nginx/html;
index index.html index.htm;
}
# redirect server error pages to the static page /50x.html
#
error_page 500 502 503 504 /50x.html;
location = /50x.html {
root /usr/share/nginx/html;
}
}
Also modified the dockerfile
FROM node:12.13-alpine as app-ui-builder
RUN npm install -g @angular/cli@8.3.14
RUN apk add --update git openssh
RUN mkdir ./app
COPY package*.json /app/
WORKDIR ./app
COPY . .
RUN npm cache clear --force && npm i
RUN ls && $(npm bin)/ng build --prod
FROM nginx:alpine AS nginx-builder
RUN apk update && apk add ca-certificates && rm -rf /var/cache/apk/*
#RUN rm -rf /etc/nginx/conf.d/*
RUN mkdir /var/logs
RUN mkdir /var/logs/nginx
COPY ./docker/nginx.conf /etc/nginx/
## Copy a new configuration file setting listen port to 8080
COPY ./docker/default.conf /etc/nginx/conf.d/
RUN rm -rf /usr/share/nginx/html/*
#
COPY --from=app-ui-builder /app/dist/app-ui
/usr/share/nginx/html
EXPOSE 8080
CMD ["nginx", "-g", " daemon off;"]
Found it via this medium post
回答3:
Does your cluster have any role Based access control Storage permissions. I also suggest that you verify Permissions required to deploy or Cloud Run for Anthos
Check if you have the Storage permission and scopes4
来源：https://stackoverflow.com/questions/59107010/gcloud-cloud-run-deployment-fails-for-deployment-to-gke