Question about bitcask disk space usage

Alexander Sicular siculars at gmail.com
Tue Mar 15 19:12:42 EDT 2011


If you are testing on a single node change either the default or the  
specific bucket nval to 1. Otherwise every write is x3.


@siculars on twitter
http://siculars.posterous.com

Sent from my iPhone

On Mar 15, 2011, at 18:53, Luke Monahan <lukemonahan at gmail.com> wrote:

> Hi All,
>
> I am just poking around riak and riaksearch at the moment, and have  
> found that the disk space usage I am seeing is about 100-to-1 for  
> the actual data size being stored. That is, for every 1MB of data I  
> am saving in Riak I am getting about 100MB of disk space used.
>
> My data has keys of variable length, but only up to about 16  
> characters. The data itself is up to about 100B only for each entry  
> (I'm saving many small entries). I'm only counting the size of the  
> bitcask directory when looking at this, starting from a clean  
> install (i.e. no overwriting of old data, which I know would simply  
> append until the cleanup process occurs).
>
> The IO levels I'm seeing when loading up data does not seem to be  
> anywhere near enough to count for the amount of disk usage. Is there  
> some thick allocation of files going on, or am I mistaken here?
>
> In any case, is the disk usage I'm seeing normal or is something  
> going wrong? I actually can't play with the data set I want on my  
> laptop with this high disk usage, so before I go and buy and  
> external SATA disk I thought I'd check to see if this is expected.
>
> Ubuntu 10.10 64-bit, Erlang and riak from source (latest release) as  
> per the wiki, ext4 filesystem, 4GB RAM. Using riak-js and nodejs,  
> both at the latest released version.
>
> Thanks in advance,
>
> Luke.
> _______________________________________________
> riak-users mailing list
> riak-users at lists.basho.com
> http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com




More information about the riak-users mailing list