Closed lgleim closed 5 years ago
I created a repo to reproduce the setup:
git clone https://git.rwth-aachen.de/lars.gleim/trellisldp.git
cd trellisldp
docker-compose up
That stacktrace tells us that Trellis received an error from Blazegraph. Can you show us what Blazegraph recorded for that request? Presumably it can log a reason for returning a 400 Bad Request response and that would be the crucial information.
My suspicion is that Blazegraph does not support the named graph format used for the audit log. If that is the case, I would suspect that updates to ACL resources would also not work. And if that is the case, there is also a straight-forward work-around on the code side.
Thanks for the pointer!
After more experimentation, I can confirm that the setup works with both Blazegraph namespaces in QUAD mode and Fuseki triplestores.
It might be helpful to add this information to the documentation in the wiki, though!
@lgleim thanks for checking into this. I added a note to the wiki under Configuring Trellis Applications, but if you have suggestions for making that more prominent, I would be open to suggestions.
I am encountering errors when trying to use trellis (current master as in the develop docker image) with a remote SPARQL endpoint. The record itself is correctly updated but the creation of the corresponding Audit Log entry fails with the following error being logged and HTTP error code 500 being returned.
My
config.yml
is as follows:The corresponding remote triplestore is backed by Blazegraph 2.1.5.
The configuration should be replicatable using the following
docker-compose.yml
andconfig.yml
at pathtrellis/config/config.yml
: