Clojurians Log v2
datalevin
👋 What does this mean exactly in the WIP changelog? > [Platform] Automatically upgrade DB from version 0.9.27 onwards. The migration may take a while when opening the DB, and it needs Internet access. How does “upgrade DB” work? Why does it need Internet access? There’s also GH issue (not from me) asking for clarification: https://github.com/juji-io/datalevin/issues/345
It needs to download the uberjar of the old version in order to dump the DB.
Just wanted to say that Datalevin is an absolute pleasure to work with. gratitude I am building an event sourced system where datalevin is "just" the nodes' local cache from which everything is served. Recently I integrated the vector DB feature as well. Having these powerful tools in one package is amazing! This is a spare-time project and I haven't hit production yet, so I cannot say how it performs under more load (but I have high hopes! 😀 )
It would be great to know how you accomplished events sourcing, as I have used it as entity store, like traditional sql db, but it's way better to work with datalevin.
My main data storage layer is really an event bus (https://docs.nats.io/). When a fresh node starts, it creates a Datalevin database on disk, creates a NATS consumer (with a newly created stable consumer ID) and reads all the events as fast as it can. Each event is as small as it can be, e.g. "update the name attribute of this entity to this value", and the processing of an event results in one or more transactions towards Datalevin. Should the node restart, it opens the Datalevin database from disk, and creates a NATS consumer with the same ID as before, such that it does not receive events it has already processed. So Datalevin is what my presentation and business logic is concerned with, while NATS is my durable data storage and event platform.
Opened a Discussion section on GitHub https://github.com/juji-io/datalevin/discussions
@huahaiy if we noticed that homebrew had version 0.9.22 and we wanted to request a bump to the latest would we do it here or under github discussion? What's the preferred avenue
Either way works for me. I checks here more frequently though. But GitHub issues are my tracking system. Will bump homebrew and docker soon.
Done