What needs to be done on an ICON orch before a shutdown of the server? I need to add more resources to the VM.
@hayden_redd, there shouldn’t be any concerns with just shutting down the server via
shutdown -h now or another command for shutdown run from the command line. The service should be gracefully stopped during the shutdown process. I believe that most hypervisors will initiate a graceful shutdown if you stop the VM, but I’m not 100% on that.
This is what I have used
For Orch Maintenance:
docker stop $(docker ps -a -q)
docker rm $(docker ps -a -q)
This will effectively clear out the old versions of plugins running. However, it won’t free all the disk spaced used. To free the disk space you’d need to run
docker image prune or
docker system prune to remove the old plugin images. If you run
docker system prune instead of
docker rm $(docker ps -a -q), you’ll remove all the plugin images. The orchestrator will then automatically re-download the ones it needs.
These are rather drastic options and I wouldn’t recommend performing them on a regular basis. Removing unused containers per Phil’s commands will clean things up, but it shouldn’t make much difference to the orchestrator since an idle plugin should just be waiting for an action to perform. If, however, a plugin container is misbehaving these commands can correct the situation BUT we’d like to make sure we know about the situation so we can fix the wayward container.
Clearing out docker can have a significant short-term negative impact on job execution times since the orchestrator has to re-download the plugins - a task it doesn’t perform until a workflow or connection test calls for executing the plugin.