Problem (again) with portsnap5.FreeBSD.org?

[RW]

On Thu, 21 Oct 2010 10:31:24 +0200
Guido Falsi <mad at madpilot.net> wrote:


> I have noticed on the machines I use/administer a bias towards
> portsnap5.
> 
> I mean, all these machines are always choosing mirror number five.
> 
> Some are behind squid proxies and using them for portsnap, so I think
> this can be expected and caused in some way by some cached data, but
> some are connecting directly.
> 
> Is this in some way expected?

If you define a cache environment variable the random choice is seeded
by that to improve the caching. There may also be some weighting in the
SRV record.