This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
2016-08-02
Channels
- # admin-announcements (3)
- # architecture (5)
- # beginners (10)
- # boot (223)
- # cider (13)
- # cljsjs (2)
- # cljsrn (50)
- # clojure (208)
- # clojure-austin (16)
- # clojure-belgium (1)
- # clojure-india (1)
- # clojure-poland (13)
- # clojure-russia (130)
- # clojure-spec (27)
- # clojure-uk (144)
- # clojurescript (135)
- # css (2)
- # cursive (10)
- # datavis (1)
- # datomic (29)
- # dirac (9)
- # funcool (2)
- # hoplon (41)
- # jobs (3)
- # leiningen (6)
- # om (37)
- # onyx (20)
- # pedestal (1)
- # planck (1)
- # proton (4)
- # re-frame (45)
- # reagent (17)
- # rethinkdb (16)
- # ring-swagger (19)
- # schema (5)
- # specter (93)
- # sql (16)
- # test-check (33)
- # untangled (7)
@shinych: from http://docs.datomic.com/storage.html#sql-database it seems that if it has a jdbc driver, you can use it, but this is software, so nothing is certain.
exactly 🙂 interested if anyone is using (has tried it) with DB2 we have kind of an enterprise hell here, and it is quite possible that DB2 would be the only option in production deployment
@shinych: you should be able to use DB2. any jdbc compliant SQL store. You’ll need to bring your own jdbc driver
you’ll also have to adapt the sql provisioning scripts included with the datomic distribution for postgres
Does transactor serializes transactions per database or is it a global queue? Is it possible to manually "shard" the DB and have parallel writes that way (ignoring the underlying storage limitations since it's all on the same DB table)?
@yonatanel: there are multiple reasons. storage throughput is one, but process isolation is another, particulars of how indexing behaves and is triggered is another
So when the license refers to 5 processes in http://www.datomic.com/pricing.html it means I can have 5 transactors?
@yonatanel: a Datomic system can only have a single transactor; the process count is the sum of your transactor and all your peers. you can have as many peers as you need/want on the system
more details here: http://docs.datomic.com/architecture.html
In one of the Nubank talks they mentioned maybe sharding/partitioning on customers. I wonder how they're going to do that.
multiple transactors. each peer can connect to multiple separate databases (and even query across them). they’ll simply need to handle traffic sharding upstream (i.e. at the load balancer or at the peer application)
in general, yes, if you’re considering needing multiple transactors we’d prefer you gave us a call so we can discuss your system and work out an appropriate licensing option
@kenny: probably not a straight up answer, but this resource should help you http://docs.datomic.com/capacity.html
you will probably adjust your transactor.properties file several times to get it just right
it’s a living system, not a “set it an forget it”
I saw that. I am talking about actual disk space, not RAM. That article has one small section about storage size but does not make any recommendations for the amount to allocate.
kennyjwilli: are you planning to run the transactor on a very storage-constrained vm or the like?
@bvulpes: No. Will be running as a marathon app in Mesos. Just wondering how much disk space I should allocate for the app.
I’m running datomic in a script (http://inlein.org): is there a way to suppress debug logging?