Hi,
first of all, nice work, I really like thredds.
We try to implement a data repository for climate data.
As data catalogue we use ckan. As webserver we use nginx with
reverse-proxies to uwsgi(ckan python application server) and
tomcat(thredds).
I wrote a small python proxy with some authorization checks to access
the thredds server. The goal is to make new subsets citable and
recallable. The WMS service works flawlessly with
https://www.nginx.com/resources/wiki/start/topics/examples/xsendfile/
so that the response is directly served via nginx instead of blocking a
worker of my application server.
For the ncss part I'd like to calculate a hash sum for the subset and
rename it before passing it back to the user. It would be nice if I
could access the cache of ncss directly from my python application to do
this. I was thinking about a job making the call to thredds, process
everything, change the Content-Disposition and return the file via
xsendfile and nginx again.
To sum up, is it possible to get a response(including the relative
filepath from the cache directory) from the thredds server when the
subset creation process is done. Can I serve the file with another stack
and wait for the deletion process due to my maxAge setting.
Is it even worth thinking about it or is it best to send the http
request from my python application server to thredds without touching
the cache.
Thank you in advance and best regards,
Georg
--
Georg Seyerl
Data Scientist / Architect
Department Data, Methods, Models
CCCA Data Centre
Central Institute for Meteorology and Geodynamics
1190 Vienna, Hohe Warte 38 (Austria/Europe)
Tel.: +43 1 36026 2512
E-Mail: georg.seyerl@xxxxxxxxxx
Homepage: www.ccca.ac.at
