Python client: Connection pool

Sean Cribbs sean at
Wed Jun 25 10:34:58 EDT 2014


No, it will only randomize the node when *opening* a connection. After
that, it just uses what is first available in the pool.

We have an outstanding plan to support auto-configuring clients, but that
work did not make it into the 2.0 cycle.

On Wed, Jun 25, 2014 at 7:18 AM, Alex De la rosa < at>

> Hi there,
> Using the Python client, in the documentation I can read this:
> "The client maintains a connection pool behind the scenes, one for each
> protocol. Connections are opened as-needed; a random node is selected when
> a new connection is requested."
> Imagine that I create my connection like this:
> client = riak.RiakClient(protocol='pbc', nodes=[{'host':'X.X.X.1',
> 'http_port':8098,'pb_port':8087},{'host':'X.X.X.2','http_port':8098,
> 'pb_port':8087},{'host':'X.X.X.3','http_port':8098,'pb_port':8087},{'host'
> :'X.X.X.4','http_port':8098,'pb_port':8087},{'host':'X.X.X.5','http_port'
> :8098,'pb_port':8087}])
> Specifying the 5 servers in my cluster... Every time that do a request
> (PUT/GET/DELETE) is going to use a randomized host from the list? Also...
> what if for example I add server X.X.X.6 and X.X.X.7? do I have to manually
> edit my connection string to reflect those changes? Is there a way to
> automatically use any node in the cluster without having to specify them
> manually?
> Thanks,
> Alex
> _______________________________________________
> riak-users mailing list
> riak-users at

Sean Cribbs <sean at>
Software Engineer
Basho Technologies, Inc.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the riak-users mailing list