Due to the current gap in continued funding from the U.S. National Science Foundation (NSF), the NSF Unidata Program Center has temporarily paused most operations. See NSF Unidata Pause in Most Operations for details.
Luca Cinquini wrote:
Hi John,is there any documentation that explains the TDS caching mechanism ? For example, are the THREDDS catalogs re-generated each time Tomcat is started, or are they serialized somewhere ?
on startup, TDS reads in all static catalogs (which can be thought of as configuration files) and caches them. If the catalog has an "expires" attribute, it will reread them upon expiration. You can force re-reading by using an expires date before todays date. You can also force rereading the catalogs through a "reinit" command, if you have https enabled and are authorized. dynamic catalogs are generated dynamically. We are considering caching them for some amount of time, but i dont think we've implemented yet (Im cc'ing ethan to double check that).
I noticed that the aggregation files are indeed cached in /content/ thredds/cache (or cacheAged).
aggregation has another set of caching going on (still evolving), and /content/thredds/cache is one of the default places it uses (can be changed). these arent catalogs, but NcML agg elements, but they are integrated into the TDS. theres some sketchy docs at: http://motherlode.ucar.edu:8080/thredds/docs/internal/Caching.htm
FYI I am sending you a diagram that shows how I enabled authorized TDS access from the CDP. We are still in prototype mode, but the idea is to start using the TDS as a FileServer, an OPeNDAP server and whatever it will do in the future.
I really want to catch up with you on this, but im swamped at the moment, so it will have to be after im back from vacation (Oct 3). If its helpful, we can talk on the phone this week, however.
thanks, Luca
thredds
archives: