There are two ways to make an application image: One is a basic image based on a packaged monitoring function (recommended), and the other is an operating system-based image (such as centos7). Just choose one.

I. Basic Image Provided by Baidu AI Cloud (Recommended)

Preliminary Preparation

1. Environmental preparation

• Open the Image Repository on the CCE product, create a namespace, and record the username and password, and the namespace name.
• A machine with Docker installed and configured.

2. Basic image

Basic image contains probes by default

Basic image description

*jdk8Installation path: /home/java *ProbeInstallation path: /home/emcprobe *start.shPath: /home

Basic image use"ENTRYPOINT"Set the container entry. When using the basic image (from), you need to use the "CMD["param1","param2"...]"format to pass the startup command for the user program. The basic image will accept the CMD command as a parameter to start the user program.

Writie Dockerfile

Take the environment based on centos7 and jdk8 Take the creation of provider-demo as an example: Dockerfile only needs to reference the basic image provided above and write its own startup command into CMD.

Applications using the monitoring function

Operation steps:

1. Go to the directory where the Docker machine stores the user program such as /home/appimage.

2. Create a Dockerfile with the following content:

   #Basic image mirror address
   FROM hub.baidubce.com/cnap-public/microservice-springcloud:probe1.0.0.1_centos7.1_java8
   # Install the application package in the /home/app/lib directory (recommended)
   COPY {BMS_JAR_NAME} /home/app/lib/
   # Application startup command
   CMD ["java","-jar","/home/app/lib/{BMS_JAR_NAME}"]

   Note: Replace {BMS_JAR_NAME} with the name of the package that needs to be deployed. The CMD command is only passed as a parameter to the basic image ENTRYPOINT command

Create Image Using Dockerfile

Run the docker build command in the directory where the Dockerfile is located (/home/appimage) to create an image:

docker build -t hub.baidubce.com/[namespace]/[ImageName]:[Image Version]. 

namespace is the name of the namespace created in the environment preparation; ImageName is the name of the image specified by the user to create;

Note: the "." after the command

When finishing the creation, you can check whether the image is created by using the docker image ls or docker images command:

docker image ls
docker images

Push the Image to the Image Library

After finishing the image production, push the produced image to Baidu AI Cloud Image Repository:

docker push hub.baidubce.com/[namespace]/[ImageName]:[Image Version] 

After the image is created and pushed, it can be deployed on the CNAP.

II. Basic Image Based on the Operating System

Preliminary preparation

1. Environmental preparation

• Open the Image Repository on the CCE product, create a namespace, and record the username and password, and the namespace name.
• A machine with Docker installed and configured.

2. Code preparation

Take the creation of provider-demo as an example:

Place the two files in the same directory of the Docker machine such as the /home/appimage directory. The following will use this directory as an example for image production.

Writie Dockerfile

2.1 Applications using the monitoring function**

Operation steps:

1. Go to the directory where the Docker machine stores user programs and Phoenix Eye probe packages such as /home/appimage.

2. Create a Dockerfile with the following content:

   FROM centos:7
   # Describe image creator information, which can be customized 
   MAINTAINER BAIDU
   # Set the time zone so that the user can use log printing to get the correct time
   RUN rm -f /etc/localtime & cp /usr/share/zoneinfo/Asia/Shanghai /etc/localtime
   RUN echo "Asia/Shanghai" > /etc/timezone
   # Install the JDK (need to manually download the jdk package and place it in the same directory as the Dockerfile) 
   ADD jdk-8u191-linux-x64.tar.gz /home/java
   # Set JAVA_HOME 
   ENV JAVA_HOME /home/java/jdk1.8.0_191
   ENV PATH $PATH:$JAVA_HOME/bin
   # Install lsof command for application diagnosis
   RUN yum update -y && yum install -y lsof
   # Set the workspace for deploying applications. Customizable. Probes are deployed in the application's flat directory
   ENV APP_HOME /home/app
   # Set the jar package name for the jar package copy. E.g. provider.jar.
   ENV JAR_NAME {BMS_JAR_NAME}
   # Install the probe package for application monitoring
   ADD emcprobe.tar.gz $APP_HOME/..
   # Install the application package
   COPY start.sh ${APP_HOME}/bin/
   COPY ${JAR_NAME} ${APP_HOME}/lib/
   # Configure the working directory
   WORKDIR ${APP_HOME}
   # Launch the application
   CMD ["/bin/bash","-c","exec ./bin/start.sh"]

   Note: Replace {BMS_JAR_NAME} with the name of the package that needs to be deployed

The application path in the Dockerfile can be customized, and the monitoring probe will be deployed in the application's flat directory. For example, the default application path is /home/app, and the monitoring probe path is /home/emcprobe. Therefore, please do not use '/' as the root directory of the application. start.sh is the application's startup script, which contains the de-registration logic and also calls the startup script of the monitoring probe. It is recommended to refer to this script to complete your own startup script writing, or use the script directly for quick access. start.sh needs to be placed in the same directory as Dockerfile. The reference of start.sh is as follows:

#!/bin/bash
# Used to log off the service in the registration center; if it is removed, the service offline may be delayed 
trap "{ stop; sleep 0.5; exit 255; }" EXIT
# For application monitoring 
export EM_APP_HOME=${APP_HOME}
export PID_FILE=${EM_APP_HOME}/log/instance.pid
cd ${EM_APP_HOME}
# Application launch logic 
function start_app(){
    mkdir -p ${EM_APP_HOME}/log/
    # Application startup parameters can be customized. You need to write pid to file after startup to use the monitoring function 
    $JAVA_HOME/bin/java -jar ${EM_APP_HOME}/lib/${JAR_NAME} &
    echo $! > ${PID_FILE}
}

# The method goes offline for specific services 
function stop(){
    echo "start to kill user process"
    if [ -f ${EM_APP_HOME}/log/instance.pid ]; then
        ps aux | awk -vpid=$(cat ${PID_FILE}) '$2==pid {print}' | grep $(cat ${PID_FILE}) &> /dev/null
        if [ $? -ne 0 ]; then
            echo "already stopped!"
            return
        fi
        echo  "stopping process "
        kill $(cat ${PID_FILE})

        for ((i=0;i<40;i++)); do
            if [ $i -eq 30 ]; then
                kill -9 $(cat ${PID_FILE})
            fi
            ps aux | awk -vpid=$(cat ${PID_FILE}) '$2==pid {print}' | grep $(cat ${PID_FILE}) &> /dev/null
            if [ $? -eq 0 ]; then
                echo -n .
                sleep 0.5
            else
                echo " [done]!"
                rm ${EM_APP_HOME}/log/instance.pid
                break
            fi
        done
    fi
}
# Launch the application 
start_app
# Start the monitoring probe 
cd ${EM_APP_HOME}/../emcprobe
bash ./bin/start.sh

Create Image Using Dockerfile

Run the docker build command in the directory where the Dockerfile is located (/home/appimage) to create an image:

docker build -t hub.baidubce.com/[namespace]/[ImageName]:[Image Version]. 

namespace is the name of the namespace created in the environment preparation; ImageName is the name of the image specified by the user to create;

Note: the "." after the command

When finishing the creation, you can check whether the image is created by using the docker image ls or docker images command:

docker image ls
docker images

Push the Image to the Image Library

After finishing the image production, push the produced image to Baidu AI Cloud Image Repository:

docker push hub.baidubce.com/[namespace]/[ImageName]:[Image Version] 

After the image is created and pushed, it can be deployed on the CNAP.

III. Historical versions

2020-7-23 2.2.0.1 latest version

• Optimize trace message generation: remove trace messages related to micro-service framework related heartbeats, pull configuration requests, and spring annotation related plug-ins
• Support brand new topology display, adapt to http and rpc call scenarios | Environment type | Description | Mirror address | | ------------------------- | ---------------------------------------------- | ------------------------------------------------------------ | | centos7.1 and jdk8 | centos7.1 pre-installed jdk8 | hub.baidubce.com/cnap-public/microservice-springcloud:probe2.2.0.1_centos7.1_java8 | | ubuntu18.04 and jdk8 | ubuntu 18.04 pre-installed jdk8 | hub.baidubce.com/cnap-public/microservice-springcloud:probe2.2.0.1_ubuntu18.04_java8 | | Monitoring probe package | Support the display of monitoring information | emcprobe-2-2-0-1.tar.gz |

2020-6-5 2.1.1.2

• Add okhttp3 plugin