Deploying our docker containers

This repository serves as a base to hold docker configuration files that can serve as examples for your own docker deployments of Panintelligence.

If you're on Azure: please see our azure-containers repository.

Requirements

Any system with docker installed or another container platform like Kubernetes.

Getting access

Please email support@panintelligence.com with the following information:

Company: Name of your company

Username: Your github username

Deployment Methods

Note None of the links to images will work unless you've been granted access.

Containers

For this method, you need the following:

• The container images:
  • panintelligence/renderer
  • panintelligence/excel-reader
  • panintelligence/scheduler (if your licence includes piReports)
  • panintelligence/pirana (if your licence includes piAnalytics)
  • panintelligence/server
• A licence from your account manager
• Access to a filesystem of some sort to persist your custom themes and the shared keys between server and scheduler

Using docker-compose

Clone this repository and edit .env according to your needs.

You may also wish to edit the docker-compose.yml To modify the load balancer config, please edit haproxy.cfg

Then stand up the service like so:

cd <code source directory>/docker-compose/example
docker-compose up -d

Volumes

for persistence or modification of certain files or features, it's well worth mounting the following volumes for your containers:

Themes

If this directory is "empty" on first load of pi, default themes will populate this directory.

please mount this directory: /var/panintelligence/Dashboard/tomcat/webapps/panMISDashboardResources/themes

Images

/var/panintelligence/Dashboard/tomcat/webapps/panMISDashboardResources/images

Custom_JDBC

Custom JDBC drivers permit you to use a freeflow jdbc config within the dashboard. Any JDBC jar file dropped into this folder will appear on the application class path and be available for use in Pi

/var/panintelligence/Dashboard/tomcat/custom_jdbc_drivers

SVG

This directory containers the custom image SVG maps.

/var/panintelligence/Dashboard/tomcat/webapps/panMISDashboardResources/svg

Logs

Currently, there's no housekeeping on the logs inside the container, so it's recommended that you volume out the logs to give space for growth. This is especially important when deploying to containers that feature a fixed root volume, such as AWS Fargate.

/var/panintelligence/Dashboard/tomcat/logs

Keys

These keys are shared with the scheduler. Its important that the same path is mounted on the scheduler container. This will allow the scheduler and dashboard application to talk with one another over an encrypted channel.

/var/panintelligence/Dashboard/keys

Further deployment configuration

For additional configuration options, see our configuration Environment Variables documentation.

Upgrading

Provided your volumes are set up as we've recommended under the Deployment Methods, upgrading is as simple as replacing the old container with a new one.

Instead of :latest you might want to use the specific version (e.g. 2023_03_30).

modify the .env tags with your new version

especially if you're using latest tags, you will want to execute a pull

cd <code source directory>/docker-compose/example
docker-compose pull
docker-compose up -d

Gotchas

• Kubernetes's own generated environment variables may clash with the some configuration environment variables, if you name your containers similarly.
• Connecting to Oracle Databases: you might need to pass in a TZ environment variable, otherwise the connection might fail with ORA-00604: error occurred at recursive SQL level 1 ORA-01882: timezone region not found. Example:--env TZ=UTC
• Azure's container instances assume the registry is elsewhere, and as such you need to prepend docker.io to the image name, i.e. docker.io/panintelligence/dashboard
• The Database must be case insensitive, i.e. the my.cnf file must have lower_case_table_names = 1 or you have supplied the switch in the command on startup to the db. an example on docker-compose.yml
  • You'll notice this if in the logs the dashboard complains about some tables being missing while the tables are in the database

mariadb-sql-injector container

We have created a mariadb sql injector container to run sql commands inside the database. This is useful if you're running on serverless environments and have restricted access to the backend.

Using docker-compose

Environment variables:

• PI_DB_SCHEMA_NAME=
• PI_DB_PASSWORD=
• PI_DB_USERNAME=
• PI_DB_HOST=
• SQL_INJECT_BOOL="true"
• SQL_FILE_NAME=

Volumes:

• :/var/sql

Getting Help

Contact support@panintelligence.com with any query or issue.