Riak Recap for Oct. 15-18

Mark Phillips mark at basho.com
Tue Oct 19 14:33:31 EDT 2010

Morning, Afternoon, Evening to all,

Here is a short recap from the last few days.



Community Manager
Basho Technologies


Riak Recap for Oct. 15-18

1)  Scott Gonyea (@acts_as) posted a repo called "Pabst." It's a
cross-platform Ruby extension for Protocol Buffers written in both
Objective-C and Objective-C++

On GitHub here ---> http://github.com/aitrus/pabst

Added to the Riak Wiki here --->

Thanks, Scott!

2) johnae and justinsheehy had a short chat about Luwak and n_val.

Gist here ---> http://gist.github.com/634452

2.1) On a related note, we have some Luwak documentation on the wiki
in case you haven't yet seen it --->

3) justinsheehy and hackeron had a brief conversation in #riak about
the differences between Riak and CouchDB.

Gist here ---> http://gist.github.com/634439

4) seancribbs, hoodos and jnewland had a great chat in #riak that
started with "Does Riak automaticly resolve link duplicates and ignore

Gist here ---> http://gist.github.com/634492

5) Q ---- Hi, I can't find this in the FAQ or with google: if i decide
to start a cluster with e.g. 64 partitions, and later want >64
machines, how big of an operation is splitting the partitions into
smaller ones? (from Tv via #riak)

    A ---- Adjusting the number of partitions after your cluster is
running not really possible ( the only way to do it would be to stand
up a new cluster and move data to it). You should take this into
account when doing your initial capacity planning for your cluster. We
generally advise that you plan for at least 10 partitions per node at
scale. In the case of 64 nodes, you might want to experimenting with a
partition count of 1024.

6) seancribbs and hoodos had a short chat that started with details on
how to list non-empty buckets and touched on some interesting use case

Gist here ---> http://gist.github.com/634547

More information about the riak-users mailing list