I work in an academic environment, and our HPC environment uses Singularity instead of Docker for containers. Singularity allows for use of Docker containers, so I'm using the official distribution on DockerHub for my pipeline. For the most part, it's a 1:1 mapping with Docker, just changing --volume to --bind for example. However, I ran into issues with loading plugins when running the docker image following this example. I'm able to get the instance running, but it doesn't load the desired plugins (APOC and GDS) into my specified directory and they aren't available in the neo4j instance. I've tried manually adding the standard whitelisting entries in the conf file, but this doesn't make the procedures available in the instance.
For now, I just downloaded the plugin jar files and placed them in the plugins folder I bind to the docker image, and this along with the whitelisting in the conf file is sufficient to get things running. Long-term, I'd really prefer to be able to specify the environment options so a user of my pipeline doesn't need to do any manual installation steps before deciding to use it in a production setting. Does anyone have any experience in using the Neo4j official docker image with plugins in Singularity?