I’ve been spinning up a few docker containers within a docker-compose.yml and recently this has begun to fail with the following error. (my transformer instance has the same issue)
Creating collector ... done
Attaching to collector
collector | Engine running in Deployment Mode
collector | Java 1.8 detected; adding $SDC_JAVA8_OPTS of "-XX:+UseConcMarkSweepGC -XX:+UseParNewGC -Djdk.nio.maxCachedBufferSize=262144" to $SDC_JAVA_OPTS
collector | bash: line 1: {ISSUES:[code:SSO_01]}: command not found
collector | Engine Pre Start failed with exit code 127
collector exited with code 127
Just executing the docker run command still works fine but sorting out the networking will be a pain so getting this functionality back would be helpful.
This used to work, but recently has stopped. I’m bringing up a local database and hopefully an instance of a Vault credential store and having the containers linked in the docker-compose file is convenient for development and testing.
The service in the docker-compose.yml file looks like this.
collector:
container_name: collector
image: streamsets/datacollector:4.3.0
environment:
- STREAMSETS_DEPLOYMENT_SCH_URL=https://na01.hub.streamsets.com
- STREAMSETS_DEPLOYMENT_ID=<my deployment id>
- STREAMSETS_DEPLOYMENT_TOKEN=<my token>
ports:
- "18630:18630"
links:
- "postgres:postgres"
As I mentioned, this had been working with 4.1.0 but has since stopped. When debugging it I could trace it to a curl command that is reaching back to streamsets.com and I suspect there’s something wrong with the command itself when run as part of docker-compose (white space missing perhaps?)
I’m running this with docker 20.10.12 on an ubuntu 20.04 system.
Thanks,
Travis
