Hi, I’m currently evaluating migrating to James for an organisation. For perspective, there are currently about 450 mailboxes using about 0.5TiB of storage, running Postfix and Dovecot with Maildirs on a Debian server.
I have some questions about the supported production configurations that didn’t quite become clear to me from the documentation. We’re aiming for a distributed deployment on our Kubernetes cluster. * Do we have to use Cassandra or is Postgres supported for a distributed setup, including JMAP? We are already running several Postgres clusters and are experienced with it, but Cassandra would be new to us, so we would prefer Postgres. * Is OpenSearch required for a distributed setup? I remember it as incredibly memory intensive, which we don’t have too much at the moment, and, well, buying more RAM is currently in nobody’s budget … * The README on Docker Hub[0] mentions `distributed`, `jpa`, `demo`, `memory`, and `cassandra` variants, for version 3.6.1. At least for 3.9.0, there is also a `postgres` variant[1] not mentioned in the README. If Postgres is feasible at all, is this image also feasible for a production deployment? * If we do switch, is mailbox migration supposed to happen via IMAP? Can we migrate while keeping IMAP UIDs and UIDVALIDITY intact, as well as folder separators, so that it will be seamless for users? What about Sieve scripts? * Are shared mailboxes/namespaces supported? Thanks, Thomas [0]: https://hub.docker.com/r/apache/james [1]: https://hub.docker.com/layers/apache/james/postgres-3.9.0/images/sha256-eb5a9bf9d7102a47ddf3927d3b5d1953a0ff82209d99a39c432c91f7c8b37858 --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
