Might be nice to include the cassandra version in the tag as many images do.

Hi everyone, have you instruct me some information to config multi nodes run in a cluster. If have a large request on 1 moment (~100-300/second - Just write/ No read), should we choose how many nodes? And how performance going with many nodes?

The spotify/cassandra:base image needs to be built before the other images. However, this ordering is not guaranteed by the Docker Hub, so new builds of the :cassandra or :cluster image usually end up using the previous version of the :base image.

The current workaround is to do a second build.

The issue is that if I'd make an image from spotify/cassandra-singlenode with modified configuration I had to modify script file which runs in entrypoint. Also the changes would be applied only when entrypoint is executed.

It looks confusing. Why singlenode configuration can't be set in image itself, i.e.
RUN sed -i .... && sed -i ...

So, entrypoint could be just a simple command like "cassandra -f".

Wdyt?

I'm trying to run the basic spotify/cassandra but I get an exception back saying unable to gossip with any seeds my host OS is ubuntu 18.04 Bionic Beaver(The host itself is installed on a VM). That's right after loading settings from cassandra.yaml

Thanks for providing this useful container.

I have a question about two cassandra configs: enable_materialized_views and enable_transient_replication. It seems the official document says "Materialized views are considered experimental and are not recommended for production use." And the default value for these two parameters are False in the document. They say they will disable them in Cassandra 4.

However, in the latest version of Cassandra (3.11.4), the default value for them are still True, as is the setting in your container's cassandra.yaml.

Shall we disable them manually for now?
Thanks!