Closed ibouziane closed 1 year ago
That'd be from https://github.com/docker-library/cassandra/blob/b3332696b4a6a4405ebf95a4c4065543ee1a073d/4.0/docker-entrypoint.sh#L43-L49
Any reason for the host networking? It'd be a lot easier without it for this image
Thank you for your answer. The reason is performance, we have better performance in this mode than when bridging, I don't know if you can confirm that too. It was possible to set the IP address via cassandra-env.sh.
If our entrypoint is causing problems, I would definitely recommend just skipping it entirely (--entrypoint cassandra
, etc).
For help doing so, I'd suggest a dedicated support forum, such as the Docker Community Forums, the Docker Community Slack, or Stack Overflow.
We are trying to run multiple containers with the same image in host mode. To every image, we affect an IP address that we configure as a sub interface. If we use the same port:
JVM_EXTRA_OPTS=-Djava.rmi.server.hostname=192.168.10.51 -Dcom.sun.management.jmxremote.rmi.port=7199 -Dcassandra.jmx.remote.port=7199
we have port already in use error. If we specify different port for each container, the host is listening on 0.0.0.0 even if we specify java.rmi.server.hostname we end up with all containers listening on all configured jmx ports and strange behavior of the cluster.