Fork me on GitHub
#xtdb
<
2021-02-15
>
tobias22:02:15

Newbie question: I'm working through the "Create an unbundled database with Crux and Confluent Cloud Kafka in 5 minutes" https://juxt.pro/blog/crux-confluent-cloud and when I try to run crux/start-clouster-node I get the error "No such var: crux/start-cluster-node". Has the API changed since that tutorial was written? If so, what's the correct function to call now?

refset00:02:48

Hi 🙂 that post on http://juxt.pro is definitely out of date now (sorry!) but it did get re-posted and freshly updated with the new API here on the http://opencrux.com site though: https://opencrux.com/blog/crux-confluent-cloud.html Out of interest, how did you stumble on to that http://juxt.pro version?

refset00:02:54

I'll make a note to update that old post, or at least place a clear note at the top of the page with a pointer to the up-to-date resource :thumbsup:

tobias00:02:47

Thanks for the link to the updated tutorial! I came across the old one via reddit (https://www.reddit.com/r/Clojure/comments/cmqhdd/crux_with_confluent_cloud_kafka_in_5_minutes/). Basically, I'm looking for the easiest way to get started with Crux for a hobby project that needs persistence. I thought that Confluent Cloud would be the way to go, but now I see that I'd need RocksDB to persist the indexes. I'm thinking that I might be better off just using RocksDB for all 3 of transaction log, document store, and index store. RocksDB just stores to disk, right, sort of like SQLite?

nivekuil01:02:34

yup, rocksdb is a kvstore on your local disk. If local disk persistence is good enough then using it for everything is the way to go

tobias01:02:57

awesome, thanks for the tips 👍

Steven Deobald01:02:31

@UA9399DFZ We're in the process of reorganizing documentation a bit. Out of curiosity, would an introductory tutorial walking you through the simplest persistent setup (just as you mentioned, 3 Rocks instances for each of the 3 stores) be something you would have liked to have seen when you were navigating the docs and blog posts?

tobias01:02:52

Yes, that would be great! I really like that you've made it so easy to get started with the in-memory db, so there's no friction to start on the "https://juxt.pro/blog/crux-tutorial-setup" tutorial. It would be great to have an obvious next step after that for when persistence is needed.

Steven Deobald02:02:28

@UA9399DFZ Thanks... I've suspected as much for a while, but it's great to hear it directly. 🙂