[Catalog-sig] [Infrastructure] PyPi "D" host outage

[Jannis Leidel]

Looks like the mirror has finally caught up.

Thanks all for letting me know :)

Jannis


On Sat, Feb 2, 2013 at 2:36 PM, Jannis Leidel <jannis at leidel.info> wrote:

> Hi all,
>
> So, I've made multiple attempts to fix the "d" mirror: I've been running
> the pep381client script manually and monitored it for 3 consecutive days.
> The simple problem seems to be a degraded connection to pypi.python.org.
> With a simple wget one of the bigger files (e.g.
> http://a.pypi.python.org/packages/source/M/MDAnalysisTests/MDAnalysisTests-0.7.6.tar.gz)
> I get 46 K/s (https://gist.github.com/bdebc7d3d55617c43d2b).
>
> Fetching the same file from *any* of the other mirrors results in ~10M/s.
> Has anything changed in the network configuration of the main PyPI that
> could cause this? Traffic shaping or the like?
>
> The IP address of the "d" mirror is 109.239.57.48 in case reaching out to
> the network staff at OSUOSL would require that information. I'm running an
> iperf server (thanks for the tip Noah) if that helps figuring out the
> problem.
>
> I'd appreciate any help figuring this out since I honestly don't know what
> else to do.
>
> Jannis
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/catalog-sig/attachments/20130204/2f93f7bc/attachment.html>