[tahoe-lafs-trac-stream] [tahoe-lafs] #2107: don't place shares on servers that already have shares
tahoe-lafs
trac at tahoe-lafs.org
Fri Nov 22 15:08:35 UTC 2013
#2107: don't place shares on servers that already have shares
---------------------------------+-----------------------------------------
Reporter: zooko | Owner:
Type: enhancement | Status: new
Priority: normal | Milestone: undecided
Component: code- | Version: 1.10.0
peerselection | Keywords: upload servers-of-happiness
Resolution: |
Launchpad Bug: |
---------------------------------+-----------------------------------------
Comment (by daira):
Replying to [comment:4 zooko]:
> [...] Here are some possible rules:
[added numbering for ease of reference]
> 1. [bad rule]: Don't put a share on a server if that server already has
a different share. Sometimes, as Daira points out, we would have to break
that rule in order to achieve Servers-of-Happiness.
>
> 2. [bad rule]: Don't put a share on a server if that share is already on
another server. The same example Daira gave makes it so we would have to
break this rule in order to achieve Servers-of-Happiness.
>
> 3. [good rule?]: Don't put a share on a server unless doing so would
increase the Servers-of-Happiness-Level.
>
> 4. [good rule?]: Don't put a share on a server unless doing so is a
necessary placement in an upload placement that would reach the required
Servers-of-Happiness-Level {{{H}}}.
I'm confused as to why we're trying to redesign this now. Doesn't the
algorithm based on Kevan's thesis that Mark has already implemented
(modulo some minor details) already satisfy 3?
Rule 4 seems ambiguous if "necessary" means "necessary to achieve
happiness". The problem is that there can be multiple ways to achieve
happiness using different subsets of a given share placement, and none of
these may be "necessary" given that the others exist.
--
Ticket URL: <https://tahoe-lafs.org/trac/tahoe-lafs/ticket/2107#comment:6>
tahoe-lafs <https://tahoe-lafs.org>
secure decentralized storage
More information about the tahoe-lafs-trac-stream
mailing list