[tahoe-dev] A simple 2 nodes replication

[Paul Rabahy]

I believe that you need to set .happy to 1. With happy set to 1, it will
still try to upload a share to both nodes, but it consider the upload a
success as long as 1 share uploaded successfully. I have found that in
practice, both shares will end up on the node that is online (when one is
offline).

Once the second node comes back online, you should run "tahoe check
--repair" to improve the redundancy of the shares that were uploaded while
a node was offline. Please verify this, I am a little unclear on if
--repair just does a repair (regenerate missing shares), or also does a
rebalance (move shares from node to node to improve redundancy).

Hopefully that helps.


On Thu, Feb 14, 2013 at 9:07 AM, thierry dijoux <tjr.dijoux at gmail.com>wrote:

> Hi all,
>
> Maybe a dumb question.... but i'm trying to set up a tiny grid (v 1.9.2 on
> a debian wheezy) with 2 nodes, but i can't have satisfaction with my setup.
> on the 2 nodes, i setup storage like this:
>
> shares.needed = 1 => as written in the doc, this is equivalent to a simple
> replication (and this is what i want)
>
> shares.total = 2
>
> shares.happy = 2 => i want a copy on each node
>
> When all nodes are up, no problems, but when a node is down, i can't
> upload anything. So just a simple question, how to achieve this (ie simple
> replication on only 2 nodes), and is it possible ?
>
> Thanks in advance for your help.
>
> _______________________________________________
> tahoe-dev mailing list
> tahoe-dev at tahoe-lafs.org
> https://tahoe-lafs.org/cgi-bin/mailman/listinfo/tahoe-dev
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://tahoe-lafs.org/pipermail/tahoe-dev/attachments/20130214/dc1a0418/attachment.html>