Hi David, We don't currently have a way of identifying particular servers or sets of servers to hold shares. Because of this, on our production network we like to control the storage servers, specifically the bandwidth to them. By design, any node can technically become a storage node, but operationally we are discouraging that for our production grid.
There are at least two volunteer grids available in which everybody participates as a storage node. One can be accessed by the following introducer: pb://[email protected]:64228,nooxie.zooko.com:64228/introducer and the other can be read about here: http://blog.okfn.org/2009/05/13/launch-of-open-data-grid/ I'm sure there are more that we don't know about :) A helper is a node which accepts encrypted data and performs the erasure coding as a service. We have several helpers in our colo and this allows users to only have to upload 1x of their data instead of 3.3x (our expansion factor for erasure coding). For slow links, this is a big help. Peter David Abrahams wrote: > on Fri May 29 2009, Peter Secor <secorp-AT-allmydata.com> wrote: > >> Hi David, >> >> First, make sure that your servers are not acting as storage nodes. >> To do this, either touch the file no_storage in the node directory, or >> in the tahoe.cfg file make sure that in the [storage] section that the >> "enabled" flag is false. > > OK. So there's no way for my own servers to participate in providing > storage for some of what I'm backing up? > >> The production introducer.furl is: >> >> pb://[email protected]:38463/introducer >> >> and the helper.furl is: >> >> pb://[email protected]:51258/ztqhnbd5srnowmax7ivn5d6wjxgjy4kc > > Thanks! Just curious: what's a "helper?" > _______________________________________________ tahoe-dev mailing list [email protected] http://allmydata.org/cgi-bin/mailman/listinfo/tahoe-dev
