It really is not an apples to apples comparison. Additionally, there was a
discussion recently about small values in Riak:
http://lists.basho.com/pipermail/riak-users_lists.basho.com/2012-October/009887.html
-Alexander Sicular
@siculars
On Oct 25, 2012, at 7:29 AM, Denis wrote:
> RIAK does no
Sure. And the default quorum of three. But I can not believe that in this
configuration, I get a speed of more than 15 000 Put / sec. Rather, it will
be about 10 000. But I'll try, if given the opportunity. A cluster of 3 nodes
with n_val = 2 I am testing at the moment - until the rate of <10 000.
FYI the documentation for Riak recommends a 5 node cluster, minimum, for
production environments for the best utilization of the redundancy and also, I
believe, load balancing across the cluster.
Chris.
> Date: Thu, 25 Oct 2012 04:29:17 -0700
> From: denis.mel...@gmail.com
> To: riak-users@list
Siculars,
David looked at the issue a bit this morning and it is caused by epmd
segfaulting on startup. Now the good news is that there isn't
something inherently wrong in Riak on OSX, but it is still not a good
thing. I suspect it has to do with OSX and it's changes to its toolchain
amongst dif
Hi Dave
Thanks for the reply. The warnings you posted on the gist do indeed match
those we see.
We're building Riak against R15B01 but the Erlang client against R15B02 as
this is embedded in our own Erlang app running on a separate server. I
don't think the errors reported against Riak and R15B02
Hello Christian -
I believe this is an inconsistency in the HTTP vs PB custom
metadata APIs. We'll be revisiting this code at some point to clean
things up. In the meantime, would you mind filing an issue at [1] so
we don't lose track of this?
Cheers -
Dave
[1] https://github.com/basho/riak-e
Hello Alexander -
I received the same issue. I'll see what I can do to get this resolved.
Cheers -
Dave
On Wed, Oct 24, 2012 at 4:47 PM, Alexander Sicular wrote:
> first time i've seen this. simply untar and start. any thoughts?
>
> npiabf100:r121 siculars$ tar xf riak-1.2.1-osx-x86_64.tar.g
Hello Martin -
Yes, riak-erlang-client 1.3.1 is the latest tag that you should be
using to build from Github. The protobuffs implementation produces
several warnings during compilation.
I attached the output of 'make' and 'rebar eunit' using Erlang R15B01
to the gist [1] referenced below. I hav
RIAK does not make sense to compare with elevelDB, objective was to assess
the cost of organizing the cluster.
I was surprised by the difference in performance. I expected to order 20 000
put / sec. Perhaps, RIAK configured incorrect?
With these results, you need at least 4 servers with RIAK to rep
Hi
We're in the process of upgrading Riak to 1.2.1 and have a question
regarding the correct stable version of the riak-erlang-client to use
within our Erlang app against this version of Riak. It seems that we should
be cloning riak-erlang-client from github and using the state of the code
as tagg
I mean, this is equivalent to...
:)
Denis
--
View this message in context:
http://riak-users.197444.n3.nabble.com/Efficiency-of-RIAK-tp4025765p4025783.html
Sent from the Riak Users mailing list archive at Nabble.com.
___
riak-users mailing list
ria
Steve, the documentation says, that yes ...
http://docs.basho.com/riak/latest/cookbooks/Benchmarking/#Configuration
===
operations
The possible operations that the driver will run, plus their “weight” or
likelihood of being run. Default is [{get,4},{put,4},{de
12 matches
Mail list logo
