jHipster App crashes propably because CloudFoundry activates cloud profile - jhipster

Question

The deployment of my small jhipster App "customerapp" fails and it is probably because cloud foundry sets the profile "cloud" in addition to the profile "dev". I am using several spaces in cloud foundry for the different stages of the development: dev, staging and prod.
I used the jhipster generator, added some entities customer, address and contacts. App is running locally without any issues.
I also use gitlab-ci to build, test and deploy my software. My .gitlab-ci.yml looks like this (I deleted some unecessary parts).
image: mydockerregistry.xxxxx.de/jutoro/jhipster_test/jhipster-dockerimage
services:
- docker:dind
cache:
key: "$CI_COMMIT_REF_NAME"
paths:
- node_modules
- .maven
before_script:
- chmod +x mvnw
- export MAVEN_USER_HOME=`pwd`/.maven
stages:
- build
- package
- deployToCF
mvn-build:
stage: build
only:
- dev
- prod
script:
- npm install
- ./mvnw compile -DskipTests -Dmaven.repo.local=$MAVEN_USER_HOME -
Dspring.profiles.active=dev
mvn-package-dev:
stage: package
only:
- dev
script:
- npm install
- ./mvnw package -Pdev -DskipTests -Dmaven.repo.local=$MAVEN_USER_HOME -Dspring.profiles.active=dev
artifacts:
paths:
- target/*.war
mvn-package-prod:
stage: package
only:
- prod
script:
- npm install
- ./mvnw package -Pprod -DskipTests -Dmaven.repo.local=$MAVEN_USER_HOME -Dspring.profiles.active=prod
artifacts:
paths:
- target/*.war
deployToCloudFoundry-dev:
image: pivotalpa/cf-cli-resource
stage: deployToCF
only:
- dev
cache:
paths:
- bin/
script:
- bash ci/scripts/deployToCloudFoundry.sh
deployToCloudFoundry-prod:
image: pivotalpa/cf-cli-resource
stage: deployToCF
only:
- prod
cache:
paths:
- bin/
script:
- bash ci/scripts/deployToCloudFoundry.sh
The DOCKERFILE (which is built and added to our docker repository also with gitlab-ci):
# DOCKER-VERSION 1.8.2
FROM openjdk:8
MAINTAINER Robert Zieschang
RUN apt-get install -y curl
# install node.js
RUN curl -sL https://deb.nodesource.com/setup_10.x | bash -
RUN apt-get install -y nodejs python g++ build-essential && \
apt-get clean && \
rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*
# install yeoman
RUN npm install -g yo
The deplpoyToCloudFoundry.sh shell script:
cf login -a $CF_API_ENDPOINT -u $CF_USER -p $CF_PASS -o "${CF_ORG^^}" -s ${CI_COMMIT_REF_NAME^^}
cf push -n $CI_PROJECT_NAME-$CI_COMMIT_REF_NAME
My manifest file:
---
applications:
- name: customerapp
memory: 1024M
#buildpack: https://github.com/cloudfoundry/java-buildpack#v3.19.2
path: target/customerapp-0.0.1-SNAPSHOT.war
services:
- postgresql
env:
#SPRING_PROFILES_ACTIVE: dev
#SPRING_PROFILES_DEFAULT: dev
#JAVA_OPTS: -Dspring.profiles.active=dev
The pipeline runs well, the app is packed into the war file and uploaded to cloud foundry as well, but it crashes and I assume it is because somehow cloud foundry still applies the profile 'cloud' and this overrides important configurations from jhipsters 'dev' profile.
[...]
2019-01-02T19:03:16.05+0100 [APP/PROC/WEB/0] OUT 2019-01-02 18:03:16.055 INFO 8 --- [ main] pertySourceApplicationContextInitializer : 'cloud' property source added
2019-01-02T19:03:16.05+0100 [APP/PROC/WEB/0] OUT 2019-01-02 18:03:16.056 INFO 8 --- [ main] nfigurationApplicationContextInitializer : Reconfiguration enabled
2019-01-02T19:03:16.06+0100 [APP/PROC/WEB/0] OUT 2019-01-02 18:03:16.064 INFO 8 --- [ main] com.jutoro.cco.CustomerappApp : The following profiles are active: cloud,dev,swagger
[...]
This later leads to:
2019-01-02T19:03:29.17+0100 [APP/PROC/WEB/0] OUT 2019-01-02 18:03:29.172 ERROR 8 --- [ main] com.jutoro.cco.CustomerappApp : You have misconfigured your application! It should not run with both the 'dev' and 'cloud' profiles at the same time.
[...]
After that cloud foundry stops the app.
2019-01-02T19:04:11.09+0100 [CELL/0] OUT Cell 83899f60-78c9-4323-8d3c-e6255086c8a7 stopping instance 74be1834-b656-4445-506c-bdfa
The generated application-dev.yml and bootstrap.yml was just modified in some places:
bootstrap.yml
uri: https://admin:${jhipster.registry.password}#url.tomy.jhipsterregistryapp/config
name: customerapp
profile: dev # profile(s) of the property source
label: config-dev
application-dev.yml
client:
service-url:
defaultZone: https://admin:${jhipster.registry.password}#url.tomy.jhipsterregistryapp/eureka/
What did I try to set the dev profile in cf:
added -Dspring.profiles.active=dev in gitlab-ci.yml in addition to -Pdev
added SPRING_PROFILES_ACTIVE: dev in the manifest env: section
added SPRING_PROFILES_DEFAULT: dev in the manifest env: section
added SPRING_APPLICATION_JSON: {"spring.cloud.dataflow.applicationProperties.stream.spring.profiles.active": "dev"} (as mentioned in https://github.com/spring-cloud/spring-cloud-dataflow/issues/2317)
added JAVA_OPTS: -Dspring.profiles.active=dev in the manifest env: section (cv env customerapp shows that it was set)
set the JAVA_OPTS -Dspring.profiles.active=dev with cf set-env and cf restage
Any help is appreciated.
Regards
Robert

Answer 1

Forget the answer before. Turns out deep down it was a datasource problem which made the app not respond to the heartbeats.
Uncomment
#hibernate.connection.provider_disables_autocommit: true
In the application properties fixed this.

Answer 2

Maybe any "future" person may stumble upon the same behaviour.
I was able to deploy my jhipster app to cloud foundry.
I somehow "fixed" it, but I am not aware of further consequences. Yet.
Turned out cloud foundry had a problem to monitor my jhipster app via the standard health-check-type http which should be "heartbeat"?
So I decided to switch the monitoring behaviour to a not heartbeat-ish way.
Just switch health-check-type to process in your manifest.yml file.
health-check-type: process
The app is now running.