Is GeoNetwork an appropriate choice when I need to automatically create large amounts of metadata?

Question

I have inherited a GeoServer installation publishes ~13,000 - 18,000 WMS layers in a dozen or so datastores on a 12 hour schedule. Other than what's in the GetCapabilities document, there is currently no metadata.

There is plenty of data scattered about, most of it is already in a format that is pretty machine readable, so the thought was to automatically create matching metadata for each layer published.

The question is: is GeoNetwork an effective platform for automating the storage & update of all this metadata, or should I:

1. write fgdc / iso xml files to the file system
2. do something else

Answer 1

Some additional info on the matter, you might want to use a WMS harvester to generate metadata from the WMS capabilities document. The geonetwork WMS harvester can create iso19115 metadata for datasets and iso19119 metadata for service.

Note that the geoserver guys are working on a CSW implementation within geoserver, which would allow you to harvest from the CSW endpoint in geoserver. https://github.com/geoserver/geoserver/tree/master/src/community/csw

Answer 2

It looks like GeoNetwork along with some python can be used for this sort of thing. This post on "Converting FGDC XML metadata records for WMS and WFS services to ISO 19139 via Python and GeoNetwork" should give a start.

GeoNetwork did at one time apparently have issues with 10's of thousands of records, but I'd hope they're fixed based by now (that advisory was two years ago).

(I appreciate it's a year later and you've probably moved on, but someone else might find this reply useful.)