[volunteergrid2-l] Impact of downed nodes.

[Christoph Langguth]

Am 06.03.2012 20:53, schrieb Shawn Willden:
> They mean that shares cannot be stored to or retrieved from the downed
> nodes, that's all.  If enough nodes holding your shares are gone, so that
> there aren't enough shares available to reconstruct your files, then your
> files are unavailable -- or lost if the nodes never come back online.
>

Another effect is that repairs will be triggered more often.
Even though I'm using 5/10/15 settings (which means that only 1/3 of the 
original fragments is required to be able to reconstruct a file), tahoe 
will trigger a repair if a single fragment is missing because some node 
is down, causing massive data transfers (See my previous mail).

Since deep-check --repair must/should be done regularly, I actually find 
this approach a bit over-eager. I'd personally think that a repair could 
be triggered more selectively (say, once half, or a third, of 
"shares.total - shares.needed" shares are missing, or so...), but... 
that's the way it's implemented.

-- Chris

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 6161 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://tahoe-lafs.org/cgi-bin/mailman/private/volunteergrid2-l/attachments/20120306/b1e649e3/attachment.bin>