Hi Michael,
I wanted to follow up on your TDS 4.3.14 remark:
We attempted to test out the latest THREDDS 4.3.14, but the NetCDF
> Subset Service included with this latest version does not work at all
> with aggregated datasets.
>
I'm running 4.3.15 (so I can't say for sure for 4.3.14) and NCSS is working
fine for my aggregates, both FMRC as well as NcML.
Examples from our testbed:
NCSS (at a point) for an NcML aggregation:
http://oos.soest.hawaii.edu/thredds-test/ncss/grid/hioos/tide_pac?accept=csv&point=true&var=ssh&latitude=20&longitude=-176&time_duration=PT24H&time_end=2013-02-15T06:00:00Z&vertCoord=0
(dataset catalog:
http://oos.soest.hawaii.edu/thredds-test/idd/tide_mod.html?dataset=tide_pac)
NCSS (at a point) for an FMRC aggregation:
http://oos.soest.hawaii.edu/thredds-test/ncss/grid/hioos/model/wav/ww3/WaveWatch_III_Global_Wave_Model_best.ncd?accept=csv&point=true&var=Thgt&latitude=31.5&longitude=171&time_duration=PT24H&time_end=2013-02-15T06:00:00Z
(dataset catalog:
http://oos.soest.hawaii.edu/thredds-test/catalog/hioos/model/wav/ww3/catalog.html?dataset=ww3_global/WaveWatch_III_Global_Wave_Model_best.ncd
)
Don't know if this is related at all, but worth a shot: I was recently
having data access problems related to aggregates in TDS 4.3 before I
updated my wmsConfig.xml file (even for services that weren't WMS, such as
ncSOS) as instructed in the upgrade notes:
http://www.unidata.ucar.edu/projects/THREDDS/tech/tds4.3/UpgradingTo4.3.html
Cheers,
John Maurer
Pacific Islands Ocean Observing System (PacIOOS)
University of Hawaii at Manoa
On Thu, Feb 14, 2013 at 10:15 AM, Michael McDonald
<mcdonald@xxxxxxxxxxxxx>wrote:
> Elizabeth,
> We are aware of this issue. Recently something has gone awry with out
> legacy tomcat/thredds server. A new set of servers are in the works!
> subscribe to forum@xxxxxxxxx for details.
>
> In the past the errors you are seeing occur when users attempt to
> query too large of a "Time range" (defaults to the entire range)
> across multiples days of data (equals the open/close of multiple large
> netcdf files in the background). I assume the NCSS was designed for
> smaller datasets, and we at HYCOM.org are pushing this to the max with
> several years of data equating to several Terabytes of netcdf objects.
> When we aggregate all of this data together into a single thredds
> catalog you should be prepared to wait a long while for the files to
> be indexed and then served.
>
> We attempted to test out the latest THREDDS 4.3.14, but the NetCDF
> Subset Service included with this latest version does not work at all
> with aggregated datasets. So the only way for this to work efficiently
> is to serve out individual files using NCSS (thus negating the
> convenience of the "Time Range" option). Will debug this in a separate
> forum thread.
>
> --
> Michael McDonald
> HYCOM.org Administrator
>
>
> > Message: 1
> > From: Joseph, Elizabeth (Perth)
> > Subject: [thredds] Downloading HYCOM using the NetCDF Subset Service
> >
> > Hi THREDDS email group
> >
> > I have been attempting to access the HYCOM data using the NetCDF subset
> service, however the website keeps timing out when I try to save the data.
> I have also limited the size of the HYCOM data in both time period
> (maximum of 2 days per download) and spatial extent in the hope that the
> website does not timeout.
> >
> > Is there a current issue with trying to download HYCOM data from the
> NetCDF Subset Service?
> > In addition, is there any possibility that the user can select which
> vertical layer they would like to download in the NetCDF Subset Service
> [e.g. I am only interested in the u and v velocity components for the
> surface layer] .
>
> _______________________________________________
> thredds mailing list
> thredds@xxxxxxxxxxxxxxxx
> For list information or to unsubscribe, visit:
> http://www.unidata.ucar.edu/mailing_lists/
>
