Create Documentation

[thataustin]

Trying to get ElephantDB running to follow along with Nathan Marz book, and I'm wondering what the best way is. Do you all just lein uberjar inside the elephantdb-server folder and run the server with java -jar [standalone filename]? That's what I'm trying to do right now, but I'm getting some arity errors. I noticed in core.clj that the -main function seems to want either of

global-config-hdfs-path: the hdfs path of global-config.clj local-config-path: the path to local-config.clj on this machine.

and merges them together. Am I even on the right path?

If someone could throw me a bone here and help me get this thing running the proper way - if there is one - I wouldn't mind writing up some docs for the README to help others get on board as well. And I'll be honest, today is my first time in Clojure world, so any tips are appreciated.

[sorenmacbeth]

erg, yeah, I've been meaning to write some better documentation forever, but I haven't gotten to it.

You have a couple of options, if your have HDFS configured and running on your EDB nodes, then you can just lein uberjar in elephantdb-server folder, adding the hadoop jars to your classpath in the manner you specified.

The way that we run it in is we create an uberjar that includes the version of hadoop we want right in the jar doing something like lein with-profile provided,uberjar uberjar, drop this jar on s3, then use a chef cookbook that we wrote to put it on our nodes.

[rucka]

Hi guys, I have problems running edb too.... I tried to compile jar edb server as you said but running local (not on s3) edb I received those errors:

---

[main] WARN c.y.metrics.reporting.JmxReporter - Error processing "fe80:0:0:0:94c7:7eff:fed2:c658%9.elephantdb":type="keyval",name="get_requests" javax.management.MalformedObjectNameException: Invalid character ':' in key part of property ... Exception in thread "main" java.lang.IllegalArgumentException: You must supply a DomainSpec when creating a DomainStore.

---

i used this configuration: -global_config.clj

{:replication 1
 :port 3578
 :hosts ["localhost"]
 :domains {
    "example" "../data/domains/example"
 }
}

-local_config.clj

{:replication 1   :port 3578   :download-rate-limit 1024   :update-interval-s 60   :local-root "./tmp"   :hosts ["localhost"]}

and I runned edb with:

java -jar elephantdb-server-0.5.2-SNAPSHOT-standalone.jar /global_config.clj local_config.clj 

Have you any suggests? Thank you

[rucka]

Hi all. I solved. I need to add a domain-spec.yalm file to the output file of domains (in my case ../data/domains/example). the domain-spec.yam file contains:

--- 
shard_scheme: elephantdb.partition.HashModScheme
persistence_opts: {}
coordinator: elephantdb.persistence.JavaBerkDB
shard_count: 32

[bmeynell]

ping @thataustin @sorenmacbeth @nathanmarz - Been perusing the Big Data book but finding it frustrating there's severely lacking information anywhere on the web describing how to install and use ElephantDB ...

[hoangtuan151]

The Big Data book of NathanMarz take me here.

Could you pls explain how to make a hours range query to ElephantDB for pageviews over time if the index is: (key, value) => key: url + "/" + granularities + "-" + bucket, value: total_views.

Does this pseudo code work: ?

key_to_search = url + "/" + granularities + "-" + "*"
elephant_db_client.find({key: {"like": key_to_search}})