Classification and Language Service (CLS) using Docker

To build a docker image and containers for Classification and Language Service:

1. Create a folder for your CLS docker project. Store all the files and downloads in this folder.

2. Download all the necessary RPM (intallation) files to needed to install the Classification and Language Service. You will need:

   • A valid licence file: Copy the licence file to project folder and name licence.

   • Java 11 x64: A supported JDK. In this example, we will be using Amazon Corretto JDK 11 RPM for Linux x64.

   • Licenced base and all language pack RPMs installation files

   • Classification Server RPM installation file

   See Semaphore component installation files for a list of RPM installation files.

3. Create a file named start-cs.sh that is comprised of the following:

   Copy

   #!/bin/bash
   
   function shutdown {
      exit 
   }
   
   echo
   echo "Starting Semaphore Classification Server on port 5058"
   /opt/semaphore/CS/bin/ClassificationServer --daemon --nouserswap 
   echo
   
   trap shutdown SIGHUP SIGINT SIGTERM
   
   while true 
   do 
      sleep 1 
   done
   

   You will use this script to launch containers. This script launches the Classification Server as a background process, and then sleeps periodically until the container stops. The trap command maps certain process signals to our custom shutdown function, which exits the script.

4. Once all the prerequisites are copied to your project folder, build the Dockerfile. The Dockerfile peforms the following tasks:

   • Starts by using the Oracle Enterprise Linux version 8 image

   • Installs all the RPMs

   • Configures some environment variables, including the SEMAPHORE_JVM_LIB variable necessary for CS language packs

   • Copies files and scripts onto the local image

   • Defines the exposed ports and /var/opt/semaphore as our external volume. The external volume at /var/opt/semaphore is very important so that you can mount a named external volume. This will ensure you do not lose logs, rulebase files and customizations each time you deploy.

   The folloiwng is an example of the docker file:

   Copy

   FROM oraclelinux:8
   
   # Copy local RPMs and licence file to image
   RUN mkdir /local-rpms/
   COPY *.rpm /local-rpms/
   
   # Get package updates and install rpms, and clean up.
   RUN yum -y --setopt=tsflags=nodocs update && \
   yum install -y --setopt=tsflags=nodocs patch && \
   yum install -y --setopt=tsflags=nodocs
   /local-rpms/java-11-amazon-corretto-devel-11.0.*.*.rpm && \
   yum install -y /local-rpms/Semaphore-LanguagePack-Linux-Base-5.6.0.rpm && \ 
   yum install -y /local-rpms/Semaphore-LanguagePack-Linux-English-5.6.0.rpm && \ 
   yum install -y /local-rpms/Semaphore-ClassificationServer-5.6.3-1.x86_64.rpm
   && \
   yum clean all # Copy licence file from local dir to image.
   COPY ./licence /opt/semaphore/licence
   COPY ./start-cs.sh /opt/semaphore/start-cs.sh
   RUN chmod 777 /opt/semaphore/start-cs.sh ENV PATH
   /usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/opt/semaphore
   ENV SEMAPHORE_JVM_LIB /etc/alternatives/jre/lib/server/libjvm.so
   EXPOSE 5058 5059
   VOLUME ["/var/opt/semaphore"]
   ENTRYPOINT ["/opt/semaphore/start-cs.sh"]
   

5. Build the image using the following command:

   Copy

   docker build -t cs563 .
   

6. Once the image is successfully built, run to build a container instance using the following command:

   Copy

   docker run -d --name cs563-1 -p 5058:5058 -p 5059:5059 \
      --mount src=sem5_cs_volume,target=/var/opt/semaphore,type=volume \
      cs563
   

   Note the use of the --mount flag. It is very important to use a named volume for the external volume indicated at /var/opt/semaphore.

   If successful, you can see this container running by using the docker ps command. For example:

   Copy

   CONTAINER ID   IMAGE   COMMAND                  CREATED          STATUS          PORTS                              NAMES
   41449c264afd   cs563   "/opt/semaphore/star..." 32 minutes ago   Up 31 minutes   0.0.0.0:5058-5059->5058-5059/tcp   cs563-1
   

7. To confirm that the Classification Server is running, navigate to the following address in a web browser:

   Copy

   http://localhost:5059
   

   Note: This value is dependent on you opening the CAT tool port of 5059, as shown with the patch command in docker file.