Friends

| -----Original Message-----
| From: Simon Peyton-Jones
| Sent: 15 September 2008 16:19
|
| You'll know that we've been having recent network problems with access to the
| GHC repository at darcs.haskell.org.
|
| I'm happy to say that Paul Heinlein, the heroic Supreme Being of Galois's
| network systems, is not only
|   (a) well aware of our difficulties, but also
|   (b) devoting effort to fixing them, and better still
|   (c) two days ago installed a box that should make things much better.

I'm happy to say that we think this problem is solve, thanks to Paul Heinlein's 
efforts. (Brief details below.)  Thank you Paul!

Yell if you see a recurrence.

Simon

-----Original Message-----
From: Paul Heinlein [mailto:[EMAIL PROTECTED]
Sent: 19 September 2008 15:02

On Fri, 19 Sep 2008, Ian Lynagh wrote:

> On Tue, Sep 16, 2008 at 06:44:39AM -0700, Paul Heinlein wrote:
>>
>> 489 KB/s (or just shy of 4000 kilobits/s) -- that's nearly wire
>> speed!
>
> I've just downloaded a stable snapshot at 479.31K/s; great stuff,
> thanks Paul!

I can't remember if I outlined for you the source of the problem. Our
local telco (Qwest) provides Ethernet over long-haul copper. The
termination unit completely confused every switch I threw at it
(Netgear, Cisco, Zytel), and the trouble ended up being a
simple-in-hindsight speed/duplex mismatch. I didn't spend much time
troubleshooting in that direction because the problem was
unidirectional: traffic move into our office much, much more quickly
than it went out. Duplex issues I've encountered in the past have
affected traffic in both directions.

Anyway, things look pretty stable at this point. Let me know if you
see any troublesome patterns in the future, but for now I'm going to
consider this problem solved.

_______________________________________________
Cvs-ghc mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/cvs-ghc

Reply via email to