Re: [thredds] Disabling the cache for file that are routinely overwritten

For #2:
In our situation, we are creating regional data files of ocean surface
currents. These currents are calculated using inputs from a network of
HF radar sites maintained by a number of institutions which are
received at our central server at various intervals. For example, at
1:00 UTC, our 0:00 UTC file is created and may only include data from
a small number of stations. Over the course of the next hour, we may
receive more data from radar stations to contribute to the 0:00 UTC
file. Typically, after 2-3 hours we have received most of the data
that is used in our file. Sometimes stations may be down for 5-10hrs,
once they come back online, those older data files arrive on our
central server and old data files are overwritten.
Thanks for bringing this up Rich!

On Fri, May 20, 2011 at 12:09 PM, John Caron <caron@xxxxxxxxxxxxxxxx> wrote:
> Hi Rich:
> 1) I wonder what happens if the TDS is in the middle of reading the file
> when it gets overwritten?
> 2) What is the use case for this overwriting?
> On 5/20/2011 12:48 PM, Rich Signell wrote:
>> TDS folks,
>> We have several providers that routinely overwrite files with new
>> data, for example, overwriting a file called "" with new
>> data each day.
>> As I understand it from reading
>> it appears that if we don't want to cache certain datasets, we have two
>> choices:
>> 1) Turn off caching for ALL datasets served by TDS, by modifying the
>> threddsConfig.xml to read:
>> <NetcdfFileCache>
>>     <maxFiles>0</maxFiles>
>> </NetcdfFileCache>
>> 2) Turn off caching for AN INDIVIDUAL dataset served by TDS, by
>> setting a<datasetRoot>  with the cache="false"/>  option, and then
>> specifying each individual dataset.
>> But if we have a bunch of datasets in a directory that we would like
>> to turn off caching for, is there a way to turn off caching in
>> <datasetScan>?
>> I tried using<datasetRoot with cache="false">  in the specification of
>> a<datasetScan>  but that didn't seem to work.
>> I certainly could have missed something.   But if not, it would be
>> nice to have more ability to control the cache on a dataset-by-dataset
>> basis in the future.
>> -Rich
> _______________________________________________
> thredds mailing list
> thredds@xxxxxxxxxxxxxxxx
> For list information or to unsubscribe,  visit:

  • 2011 messages navigation, sorted by:
    1. Thread
    2. Subject
    3. Author
    4. Date
    5. ↑ Table Of Contents
  • Search the thredds archives: