For some reason, the postgres plugin segfaults when I set EnableSsl to true. I know the rest of my config is good, because everything works as long as EnableSsl is set to false. And I know my postgres instance is good, because I’ve been able to connect to it over SSL using pg_admin. Currently using the jodogne/orthanc-plugins:latest docker image. My postgres instance is hosted in AWS, using RDS.
For future reference, here is how to start a Docker container running PostgreSQL with SSL enabled onto local TCP port 5000 (the Docker image includes a self-signed certificate for test purpose):
Thank you so much, Sébastien! I’m now using the osimis/orthanc image, and the SSL connection to Postgres is working great!
I am curious, though – how should I make the decision of which docker image to use? Right now it seems like a good idea to use osimis/orthanc, since it works with Postgres SSL, and it also allows me to configure Orthanc using environment variables. Is it generally recommended that people use the osimis/orthanc image? Or are there situations where someone would want to use the jodogne/orthanc image instead?
Summarizing, the “jodogne/orthanc” image is used by the core developer of Orthanc to test the cross-distribution binaries (compiled using LSB), and to test the raw configuration file of Orthanc. This is a “dev” image. The “osimis/orthanc” is typically aimed at “ops” teams.
In either case, the LSB binaries should not crash with PostgreSQL+SSL: This is a bug for which there is a known workaround (use “osimis/orthanc”).
The problem is that I’m currently overwhelmed with way too many requests (this forum has received 353 messages since January 2021), so I can’t fix such an issue by now. I have introduced a bug to keep track of this: https://bugs.orthanc-server.com/show_bug.cgi?id=193