On Sunday 21 October 2001 07:33, Oskar wrote:
> On Sun, Oct 21, 2001 at 10:58:15AM +0100, Adam Langley wrote:
> > On Sat, Oct 20, 2001 at 05:29:39PM -0700, Ian Clarke wrote:
> > > I can't imagine why it would abort a request just because one node
> > > cleanly rejects it.
> >
> > Well, if a TimedOut counts as cleanly rejected - then it's ok.
>
> No. Most probably it is simply that Ian's node only has one working ref,
> so when that rejects the request (cleanly) it has nowhere else to go.
>
> > AGL
> >
> > --
> > 90% of generation[x] will always think that generation[x+2] are too
> > liberal.

This problem is probably the result of my last check in.

I was trying to implement CP only contact management as Oskar suggested.  It 
seems to work well once the node is well integrated into the network but can
execessively punish noderefs at startup, completely disabling routing.

I think that the right thing to do at this point is to put the CP failure 
recovery code back in.

--gj



-- 
Freesites
(0.3) freenet:MSK at SSK@enI8YFo3gj8UVh-Au0HpKMftf6QQAgE/homepage//
(0.4) freenet:SSK at npfV5XQijFkF6sXZvuO0o~kG4wEPAgM/homepage//

_______________________________________________
Devl mailing list
Devl at freenetproject.org
http://lists.freenetproject.org/mailman/listinfo/devl

Reply via email to