CSW Harvest in 2.4.2 issues

[ddnebert]

We set up an instance of the server attempted to register two remote catalogs for direct and full periodic harvesting. One is the same version of GeoNetwork, the other is a CSW 2.0.2 solution which is also queryable by CSW externally. The harvest of the other GeoNetwork node appears to work, though there is no feature to select all records, to issue a Null Query (standard practice to get all), or to specify a global BBOX - another trick to fetching all records.

For the CSW 2.0.2 (ebRIM Profile and baseline), I included the service URL: ​http://geossregistries.info:1090/GEOSSCSW202/discovery?Request=GetCapabilities&Service=CSW&Version=2.0.2 , two search criteria blocks "component" and "service" (without quotes). I set Groups to All and selected all three checkboxes and selected other information resources as the category.

I selected "Run" and get a java error "raised exception when searching" when I hover over the warning icon. This is a less-than-helpful diagnostic. How do I successfully harvest this CSW?

[ddnebert]

Replying to ddnebert:

Here is the console output of the CSW harvest error:

912252 [Thread-122] WARN geonetwork.harvester - Raised exception when searching

: org.jdom.input.JDOMParseException: Error on line 6: The element type "link" m

ust be terminated by the matching end-tag "</link>". 912252 [Thread-122] WARN geonetwork.harvester - Raised exception while harvesti ng from : Air Quality Community CSW (CswHarvester) 912252 [Thread-122] WARN geonetwork.harvester - (C) Class : OperationAborted Ex 912252 [Thread-122] WARN geonetwork.harvester - (C) Message : Raised exception

when searching

OperationAbortedEx : Raised exception when searching

at org.fao.geonet.kernel.harvest.harvester.csw.Harvester.doSearch(Harves

ter.java:397)

at org.fao.geonet.kernel.harvest.harvester.csw.Harvester.search(Harveste

r.java:223)

at org.fao.geonet.kernel.harvest.harvester.csw.Harvester.harvest(Harvest

er.java:94)

at org.fao.geonet.kernel.harvest.harvester.csw.CswHarvester.doHarvest(Cs

wHarvester.java:211)

at org.fao.geonet.kernel.harvest.harvester.AbstractHarvester.harvest(Abs

tractHarvester.java:379)

at org.fao.geonet.kernel.harvest.harvester.Executor.run(Executor.java:87

)

[josegar]

The ​http://geossregistries.info:1090/GEOSSCSW202/discovery?Request=GetCapabilities&Service=CSW&Version=2.0.2 server seem having some issues:

• When requested summary records returns a full record (fixed in harvester to manage full record)

• Seem always returning all records independent of filter. For example:

<ogc:Filter xmlns:ogc="http://www.opengis.net/ogc">

<ogc:PropertyIsEqualTo>

<ogc:PropertyName>dc:title</ogc:PropertyName> <ogc:Literal>non valid title</ogc:Literal>

</ogc:PropertyIsEqualTo>

</ogc:Filter>

returns 348 matches. Any value used in title returns same number of records.

• The metadata returned is not correctly escaped: some metadata contains unescaped characters like '&' in abstract, what is causing errors in saxon

Summarizing, fixed Harvester.getRecordInfo to manage full record if returned by the server instead of summary record. The rest seem a problem of this CSW server.