Clojurians Log v2

Morning all. Yesterday's meetup went wrong direction for me. I finished 3 exercises but didn't get much of feedback. Need to put more focus on feedback next time :simple_smile:

I didn't spot structure in Bob so I just brute-forced it :simple_smile: http://exercism.io/submissions/072d4cb6789f44a68627ec11f8b11646

Rna Transcription http://exercism.io/submissions/a2df77d02a6d48618e9008ed67fa9a38

Word Count http://exercism.io/submissions/e93db133d3c847bcab632ec998b36e60

@lucasbradstreet: Are there any plans to write comparison of Onyx against competition? something like Samza team did. https://samza.apache.org/learn/documentation/latest/comparisons/storm.html https://samza.apache.org/learn/documentation/latest/comparisons/spark-streaming.html

We’ll definitely need something like that in the future

@lucasbradstreet: congrats on onyx looks pretty cool. One question though.. how big are the onyx logs stored by zookeeper? By “log” do you have the same meaning as say a kafka log?

They’re only used for cluster coordination, so they’re pretty small. Stuff like - peer joined cluster, peer left cluster, job submitted.

ok cool so basic coordination stuff… that’s good, ZK doesn’t like big nodes :simple_smile:

It's also GC'able

Yup :)

may want to revise this statement: "Rather, peers in Onyx coordinate via a shared, immutable, log that is written to ZooKeeper"

makes it sound like a journal is being stored to ZK

actually… i guess the journal is command messages

not actual data

is that correct?

Which is true, it's just a small journal. Sounds like it could use some improvement still.

Yep that's right.

Data is communicated peer to peer

gotcha :thumbsup:

I'm off to sleep, but I'm happy to chat about it as much as you like another time.

no probs… take care

TIL: you can define pre and post conditions for your clojure functions http://blog.fogus.me/2009/12/21/clojures-pre-and-post/