Mirror bandwidth and user redirection

[Patrick O'Callaghan]

On Mon, 2008-06-30 at 14:25 -0500, Matt Domsch wrote:
> On Mon, Jun 30, 2008 at 12:23:08PM -0430, Patrick O'Callaghan wrote:
> > On Mon, 2008-06-30 at 10:47 -0500, Matt Domsch wrote:
> > > On Mon, Jun 30, 2008 at 11:12:45PM +0800, Ed Greshko wrote:
> > > > Patrick O'Callaghan wrote:
> > > > >On Sun, 2008-06-29 at 23:46 -0500, Matt Domsch wrote:
> > > > >>The basic selection algorithm for choosing
> > > > >>the order in which to return mirrors to clients remains the same:
> > > > >>prefer same netblocks, internet2 in same country if on internet2, same
> > > > >>country, same continent, then global, in that order.
> > > > >
> > > > >That's totally logical, but it's wrong for some cases. Here in Venezuela
> > > > >there is much better bandwidth to the US than to anywhere else in South
> > > > >America, so the "same continent" rule is not going to work for us. I
> > > > >suspect the same is true for some other SA countries.
> > > 
> > > Understood.  But I don't have a way to know that.
> > 
> > Of course. What's needed is a way to tune these things manually.
> 
> append "&country=us,ca,mx"  to the end of your mirrorlist URLs listing
> any countries you think would be faster for you.

OK, I'll try that.

> > My only suggestion for now is that the weighting of the various classes
> > be changeable via a config file. I don't know if that is easy or hard to
> > do given the existing code.
> 
> hard.  The whole point is to have a system that doesn't require user
> config file changes, but that is "good enough" for nearly everyone.  A
> few items can be changed, like appending &country=  or &ip=  to
> override the normal detection mechanisms, but I don't want to make it
> infinitely configurable by users.  &country=global works too.

"Allow" isn't the same as "require", but I understand your point.

> 
> > > > >Also, for the relatively few people on Internet2 it's always better than
> > > > >Internet1, at least here. I mean Internet2 to anywhere is better than
> > > > >Internet1 to the same city.
> > > 
> > > That all depends on the interconnects between the nodes on Internet2
> > > and the commerical internet.  As those links cost real money for our
> > > volunteer mirror admins, by request of some of the I2 mirrors in our
> > > system, I've tried to avoid sending non-Internet2 users to Internet2
> > > servers.
> > 
> > That's fine. I'm talking about I2<->I2 connections, which if available
> > should outweigh non I2<->I2 connections even if the latter are more
> > local.
> 
> Here I restrict it to I2<->I2 within the same country, as I don't know
> the I2 connectivity between countries.  Maybe it's faster, maybe
> not...

I2 is intended to support high bandwidth connections that don't congest
with normal Internet traffic. Here, it's definitely faster, but YMMV of
course.

> Users can always use yum-fastestmirror if they like.  That has the
> advantage of using the mirrorlist, but with timed values from the
> actual client.

The problem with yum-fastestmirror is that it only measures latency, not
bandwidth. Also, it appears to ping every server on its list every time
it runs (even though it keeps a cache file of results) so I suspect it's
buggy.

poc