I would second this request. It would give users who have been
scrambling this week some time to do a full test and for Unidata to
resolve any issues that arise during testing (like my broken GFS bundle).
Don
On 4/2/14 1:27 PM, Tyle, Kevin R wrote:
Given the problems that arose with the first rollout last Friday, I think at
least an extra week of testing is necessary. I.e. the earliest it would be
rolled out should be no earlier than Monday, 14 April.
_____________________________________________
Kevin Tyle, Systems Administrator
Dept. of Atmospheric & Environmental Sciences
University at Albany
Earth Science 235, 1400 Washington Avenue
Albany, NY 12222
Email: ktyle@xxxxxxxxxx
Phone: 518-442-4578
_____________________________________________
-----Original Message-----
From: idvusers-bounces@xxxxxxxxxxxxxxxx
[mailto:idvusers-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Don Murray
Sent: Wednesday, April 02, 2014 3:22 PM
To: Sean Arms; THREDDS community; idvusers@xxxxxxxxxxxxxxxx
Subject: Re: [idvusers] [thredds] Update - Changes to thredds.ucar.edu
Sean-
On 4/2/14 11:17 AM, Sean Arms wrote:
We were thinking of proposing a testing period of this week, through
the weekend, and see where things stand on Monday. Given that the test
plugin makes the process of testing bundles no more difficult than
running the bundle itself, I would think that would be enough time. We
will send an email to the different users of THREDDS with specific
changes and instructions on how to best test.
What is your solution to the variable name changes that I encountered in the
IDV for my bundle? Since the mapping of old to new names is contained in the
IDV through the calls to the netCDF-Java library, this would require a new IDV
release. But the current IDV used ncIDV 4.3, not 4.5 and it's unrealistic to
require everyone to upgrade to a new version on such short notice.
> Please let me know when you have this working in the IDV. ;-)
I will look into it. As you know, this has been an issue with the IDV
for many, many years now, and this particular suggestion is not
something new. But I'll be sure to let you know how things go ;-)
Actually, this is the first time I've heard anyone bring up this suggestion on
the data access. Each request for data for a timestep is a separate server
request, so I'm wondering what one would do during a multithreadded loading
when the server goes down in the middle of loading a long set of timesteps.
Hopefully, you'll figure it out.
Don
--
Don Murray
NOAA/ESRL/PSD and CIRES
303-497-3596
http://www.esrl.noaa.gov/psd/people/don.murray/
