Would it be useful to include this information to the backup docs for users
who are now using this version without the UI, Andy? I can create a PR for
that if needed.

Cheers
Bruno


On Mon, 28 Nov 2022 at 10:38, Andy Seaborne <a...@apache.org> wrote:

>
>
> On 27/11/2022 21:21, Tim McIver wrote:
> > Thanks Andy.  I probably won't have a large amount of data for some time
> > but I can imagine getting a timeout or other error if there is a huge
> > amount of data.  Can it be returned compressed?
>
> That can be done with
>
> curl http://<fuseki-host>:3030/ds | gzip > data.trig.gz
>
> and there won't be a client-side whole uncompressed copy at any point.
>
>      Andy
>
> >
> > Tim
> >
> > On 11/27/22 13:53, Andy Seaborne wrote:
> >>
> >>
> >> On 27/11/2022 18:28, Tim McIver wrote:
> >>> Thanks Bruno!  Yeah, that tracks since I know that LinkedDataHub
> >>> <https://github.com/AtomGraph/LinkedDataHub>
> >>
> >> and to be clear - provided by AtomGraph, not the Apache Jena project.
> >>
> >>  uses the "no UI" version
> >>> <https://hub.docker.com/r/atomgraph/fuseki> but I hadn't realized
> >>> that it did not come with the admin functions.
> >>>
> >>> So a follow up question would be: can I still somehow back up the data?
> >>
> >> The backup is an RDF file
> >>
> >> Another way to get an RDF file of all the data is
> >>
> >>    curl http://<fuseki-host>:3030/ds
> >>
> >> will return the data as well. It's just as valid and is a isolation
> >> snapshot.
> >>
> >> This is the simplest way.
> >>
> >> The backup operation, if it were available, schedules a background
> >> task. That task writes to a local file which you can copy after the
> >> backup has finished.
> >>
> >>>
> >>> I have access to the data files from Fuseki and I considered just
> >>> copying or rsyncing them but I know that, in general, that's not safe
> >>> as I can't expect them to /not/ be updated while I'm backing them up.
> >>
> >> Correct - copying the database files while it is running is, in
> >> general, not safe.
> >>
> >> If the database is compacted, the old generation data area is safe to
> >> copy.
> >>
> >>     Andy
> >>
> >>>
> >>> Tim
> >>>
> >>>
> >>> On 11/27/22 13:18, Bruno Kinoshita wrote:
> >>>> Ah, at the top of this page:
> >>>>
> https://jena.apache.org/documentation/fuseki2/fuseki-main#fuseki-docker
> >>>>
> >>>> It says: "Fuseki main is a packaging of Fuseki as a triple store
> >>>> without a
> >>>> UI for administration." And further down: "The main server does not
> >>>> depend
> >>>> on any files on disk (other than for databases provided by the
> >>>> application), and does not provide the Fuseki UI or admins functions
> to
> >>>> create dataset via HTTP.". I had forgotten about that change.
> >>>>
> >>>> So I believe you are right Tim, what you must have in your container
> is
> >>>> Fuseki main without the UI, so without the backup servlet & endpoint
> >>>> binding (thus the 404). You can have a look at the page about Fuseki
> >>>> + UI
> >>>> for options for running it separately with access to admin features:
> >>>> https://jena.apache.org/documentation/fuseki2/fuseki-webapp.html
> >>>>
> >>>> -Bruno
> >>>>
> >>>> On Sun, 27 Nov 2022 at 19:12, Bruno Kinoshita<brunodepau...@gmail.com
> >
> >>>> wrote:
> >>>>
> >>>>> I got the same result following the docs for the Docker compose
> >>>>> installation:
> >>>>>
> https://jena.apache.org/documentation/fuseki2/fuseki-main#fuseki-docker
> >>>>>
> >>>>> Adding --update didn't solve it. So there might be something that
> >>>>> needs to
> >>>>> be enabled in the dataset assembler configuration when you create the
> >>>>> dataset in the container.
> >>>>>
> >>>>> On Sun, 27 Nov 2022 at 18:56, Tim McIver<t...@timmciver.com>  wrote:
> >>>>>
> >>>>>> It's not working for me.  I even tried doing it from the fuseki
> >>>>>> container.  It seems this image does not have curl so I tried wget
> >>>>>> using
> >>>>>> 'wgethttp://localhost:3030/$/backup/ds  --post-data ""'. Again, I
> >>>>>> get a
> >>>>>> 404.
> >>>>>>
> >>>>>>
> >>>>>> Do the admin endpoints have to be specifically enabled? Or could
> they
> >>>>>> have been disabled?
> >>>>>>
> >>>>>> Tim
> >>>>>>
> >>>>>> On 11/27/22 12:07, Bruno Kinoshita wrote:
> >>>>>>> Hi Tim,
> >>>>>>>
> >>>>>>> I am not using a container, but I just tested the latest version
> >>>>>>> from
> >>>>>> Git
> >>>>>>> on Eclipse, and tested the endpoints with curl to query and backup.
> >>>>>> Maybe
> >>>>>>> your endpoint URL is missing something?
> >>>>>>>
> >>>>>>> 1. Create ds in-memory dataset
> >>>>>>> 2. Load some dummy data
> >>>>>>> 3. curl a query: $ curl 'http://localhost:3030/ds/' -X POST
> >>>>>>> --data-raw
> >>>>>>> 'query=...' (success, data returned as expected)
> >>>>>>> 4. curl to trigger a backup: $ curl
> >>>>>>> 'http://localhost:3030/$/backup/ds'
> >>>>>> -X
> >>>>>>> POST
> >>>>>>>
> >>>>>>> Then, if you want, you can also query for the tasks (a back up
> >>>>>>> creates
> >>>>>> an
> >>>>>>> async task on the server):
> >>>>>>>
> >>>>>>> $ curlhttp://localhost:3030/$/tasks
> >>>>>>> [ {
> >>>>>>> "task" : "Backup" ,
> >>>>>>> "taskId" : "1" ,
> >>>>>>> "started" : "2022-11-27T18:06:01.868+01:00" ,
> >>>>>>> "finished" : "2022-11-27T18:06:01.893+01:00" ,
> >>>>>>> "success" : true
> >>>>>>> }
> >>>>>>> ]
> >>>>>>>
> >>>>>>> -Bruno
> >>>>>>>
> >>>>>>> On Sun, 27 Nov 2022 at 17:55, Tim McIver<t...@timmciver.com>
> wrote:
> >>>>>>>
> >>>>>>>> I should mention also that the Docker image that I'm using in
> >>>>>>>> this case
> >>>>>>>> comes from here<https://github.com/AtomGraph/fuseki-docker>.
> >>>>>>>>
> >>>>>>>> On 11/27/22 11:43, Tim McIver wrote:
> >>>>>>>>> Hello,
> >>>>>>>>>
> >>>>>>>>> I'd like to backup my Fuseki data using the web API. I found
> >>>>>>>>> documentation about how to do that here
> >>>>>>>>> <
> >>>>>>
> https://jena.apache.org/documentation/fuseki2/fuseki-server-protocol.html#backup
> >>>>>>> .
> >>>>>>>>> But when I try use the listed endpoints, they all result in a
> 404.
> >>>>>>>>> I'm using curl from a container in a Docker network to do this. I
> >>>>>>>>> know that I can connect to the server because a call like "curl
> >>>>>>>>> http:/<fuseki-host>:3030/ds" returns data with content type
> >>>>>>>>> "application/trig".
> >>>>>>>>>
> >>>>>>>>> What am I missing? Any help would be appreciated.
> >>>>>>>>>
> >>>>>>>>> Tim
> >>>>>>>>>
>

Reply via email to