Re: [Help-gnunet] INDIRECTION_TABLE_SIZE and download speed

[Christian Grothoff]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Sunday 08 September 2002 02:15 pm, you wrote:
> On Sun, Sep 08, 2002 at 01:27:04AM -0700, Tracy R Reed spake thusly:
> > Also, I have two nodes running. I inserted a bunch of content on one node
> > and searched for it on the other node. Both nodes were well connected
> > with 10 and 13 hosts connected each. It didn't find it even after a very
> > long time. With only about 20 nodes in the network how can this be?
>
> Shortly after I sent the above email I created a 5 meg file of
> stuff from /dev/urandom on each of my nodes and inserted them. Within a
> minute they were both able to find and download each others files. So I
> must have just had bad luck the first time.

Well, some of the 'random(m|n)ess' in GNUnet will certainly create the 
momentum of luck, but unless there is a bug, I would hope that you should be 
lucky pretty much all the time :-{)

> I have noticed another interesting problem: During my unsuccessful
> attempts to downlod files I ended up with a lot of corrupted files but
> with the proper size in my directory of files to share which then got
> reinserted when I ran a script to insert my own content, contributing to
> the spread of broken files. I think the download client should put files
> in another tmp directory until they are 100% downloaded and then moved
> somewhere else to indicate they are complete so there is no chance of
> incomplete files accidentally being indexed.

Well, files are always downloaded to where the user specifies they should go. 
The idea why I took the final location also came from the fact that if you 
resume a download later (by re-issueing the gnunet-download command), we 
would use the 'corrupted' file as a starting point -- that's also why it is 
not removed. 
Any other opinions on this? Hmm. Maybe we should start putting 'GNUnet polls' 
on the webpage for issues that can be resolved in many ways and where it's 
more a question of what users want and less of a technical issue. What do you 
think?

Christian
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE9e72J9tNtMeXQLkIRArS9AKCiyNn7rtoDUJ6vz/1uuXgfykn/PgCeLbvC
zqpD/aX1OvmNj7fwiL+cqNE=
=hI9t
-----END PGP SIGNATURE-----