[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: occasional weird timestamps and bad data with THREDDS

This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.

  • Subject: Re: occasional weird timestamps and bad data with THREDDS
  • Date: Wed, 30 Mar 2011 12:12:48 -0600
Hi Tom:

1) what version TDS?
2) is this an aggregaton? (you can send me the config catalog)
3) are there errors in threddsServlet.log?
4) when you "update the netcdf files multiple times over 5-6 hours ", do you overwrite? extend?
its hard when its intermittent, we may have to figure out how to consistently reproduce. 

john

On 3/30/2011 11:21 AM, tom cook wrote:

HI John,
I have been tasked with maintaining the HF Radar national network
THREDDS server here at SIO. Rich has helped us set things up, and has
been a valuable resource for learning this system. We have been having
a frequent problem that Rich has been unable to help us with, which is
occasional bad timestamps and occasional bad data served through
THREDDS.
An example of bad data being served through THREDDS: we are working
with ASA&  USCG to get data into their SAROPS product. They have
complained that many of the data samples have had data in the wrong
geographical location (ie ocean currents over land). I have noticed
this as well using the Godiva viewer, and when I compare the Godiva
map with the source netcdf file, it is not plotting the data in the
correct location. This makes me think it is a problem with the data
being served by THREDDS. The problem goes away after some time, as we
typically update the netcdf files multiple times over 5-6 hours after
they are created. I'm not sure if the updating of the netcdf files is
the cause of the problem, but it is not something that is easy to
reproduce.

The problem we have with weird timestamps happens occasionally, and in
random locations throughout the time index. It usually is just one
sample here or there, and not successive samples. It almost always
goes away when I empty the cache.

I'm not sure if these issues are setup related, or issues we can
script some workarounds to help. But I wanted to check with you to see
if you have any idea of what is going on.

Thanks for your time,

Tom Cook
Programmer/Analyst
Coastal Observation Research&  Development Center
Scripps Institution of Oceanography