Hello Members,

Going into the weekend...

We expect peak times on Saturday and Sunday to cause slowness and
timeouts as even though our filesystem had a good week, demand is
heavy on the weekend.  Between 18:00 and 02:00 GMT is expected to be
the hardest hit.  We think this weekend won't be as bad as prior
weekends, however.

-Slow downloads.
-Connection timeouts or disconnections.
-Missing content.

Workarounds:

-Contact Support with the URLs of the timeouts.
-You can use the Beta NNTP farms. (No longer Beta after Monday)
-You can schedule downloads to occur between 06:00 and 14:00 GMT when
our system load is the least.  Weekdays are lower than weekends.

-We've armed Support with a few more tools that should help them turn
around some of the more common timeout issues.  If you experience
timeouts, send the URLs (only up to the end of the hash) to Support
and it will allow them to collect data on where the slow point in the
filesystem is and potentially workaround it shortly after the report.

-We have 25 new fileservers racked, powered, networked, configured and
in testing that I since haven't mentioned.  These aren't our big guns,
(If you like EVNova, think Valkyrie, not The Unrelenting)
and don't represent any of the new 250,000 IOPS, but should begin to
help with the bottlenecks as soon as Saturday, but more likely Monday
due to continued burn in testing.

-About 20% of our new gear is now powered awaiting some final fiber
connections.  We're still burning this gear in.

-The remaining 80% (of the over 250,000 IOPS and 500 TB we're adding)
is currently shipping.  Deliveries are expected starting Monday
into next week.

-We're decomissioning the old news.easynews.com NNTP gear to make room
for the new web gear.  We're making the Beta NNTP sites live next
Monday.

-App Status is almost complete and I was too optimistic when I wrote
it'd be released by today - but use it and give us feedback.  Support
will keep it updated with confirmed issues.
http://www.easynews.com/appstatus/

We know we can stabilize the platform, scale it, and engineer ways
forward to even better service, but that doesn't mean we won't hit
bumps in the road.  We're putting in a
few 'early warning systems' that will lower the liklihood of such
issues in the future.  As always, we rely on knowledgable members
posting in the forums and reports of strangeness from all of you.

We really appreciate your continued support!

Thanks,
Boromir

---------------