If you really want an S3 API layer around NFS, I suspect you'd be better off taking the code from Riak CS and creating your own solution than attempting this approach.
One example of how this would run into trouble: if any server fails, nearly every large object would be unavailable because of the N=1 replication factor. The data itself would be available via NFS, but the server that's supposed to serve up those chunks wouldn't. -John On Oct 21, 2013, at 9:08 AM, Adrian Nicolae <adrian.nico...@constanta.rdsnet.ro> wrote: > > Hello, > > I want to try the following setup for a cloud-storage project : > > - a cluster of x86 servers with SAS drives running Riak CS > > - the replication for the metadata will be Nx3 on the servers running Riak > > - data for the objects will be stored on a NFS shared storage (SONAS > probably) with no replication configured on Riak (Nx1). The NFS platform will > handle data protection instead in the backend. > > I found a similar setup in MezeoCloud. > > What do you think about this ? > > > Thanks, > > Adrian. > > _______________________________________________ > riak-users mailing list > riak-users@lists.basho.com > http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com _______________________________________________ riak-users mailing list riak-users@lists.basho.com http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
