Hi Pietro,

What mechanism were you using to upload the file?

A couple of ways to reload:


   1. Rebuild the database offline with tdbloader, stop the server,
   replace the database directory with the new one and restart.
   2. Load via the UI - this might be a problem for TDB1, depending on heap
   space but TDB2 does not have the same size of transaction limitations.
   3. Use curl to PUT (or POST - ad the data, not removing the old data) to
   the dataset URL.


Compression sizes: my rule of thumb is x8-x10 compression of N-Triples, a
bit more for N-Quads with a small number of named graphs. But it is all
deata shape dependent. If there are a lot of literals the same, the
compression can be much higher.

Andy


On Mon, 17 Feb 2020 at 09:13, Pietro Liuzzo <[email protected]> wrote:

> Dear Andy,
>
> Thank you very much!
>
> That was very helpful, I managed to get those backups.
>
> Now however I face a new issue. What is the recommended way of restoring?
> I tried to load the backups from Fuseki using upload files. There I tried
> to load the backup file as it was, the unpacked .nq, and even bits of it,
> but it looks like with more than 2Mb I already get Request Entity Too
> large. Should I set this limit somewhere? The backup file is only 2mb, the
> unpacked one is 50mb.
>
> Thank you very much for your support
>
> All best
> Pietro
>
>
>
> Pietro Liuzzo
> ORCID: https://orcid.org/0000-0001-5714-4011
>  https://www.aai.uni-hamburg.de/en/ethiostudies/study
>
> > Am 15.02.2020 um 13:42 schrieb Andy Seaborne <[email protected]>:
> >
> > Hi Pietro,
> >
> > On 15/02/2020 05:22, Pietro Liuzzo wrote:
> >> Dear all,
> >> I am moving to a new server, and I would like to backup the data in my
> graphs from the old instance to restore them in the new one.
> >> In the running Apache Jena Fuseki I have hit the backup button, which
> informs me it has started and finished but I am not sure where to find now
> this backup so that I can use it to restore data in the new instance.
> >> I have tried to access /etc/fuseki/backups but it looks like I cannot.
> >
> > The directory will be owned by the id that is running the server.  You
> may want to change the access permissions.
> >
> >> Can I change in my configuration the location where the backups are
> stored?
> >
> > Only by using symbolic links - the backups go to a fixed location; that
> can be a symbolic link to somewhere else.
> >
> > Different approach: access the data via HTTP:
> >
> > curl --header 'Accept: application/trig' \
> >    http://localhost:3030/ds > myData.trig
> >
> > which does an HTTP GET on the dataset, asking for TriG format.  This is
> available when there is GSP (SPARQL Graph Store Protocol) functionality.
> >
> >    Andy
> >
> >> Thank you very much,
> >> All best
> >> Pietro
> >> Pietro Maria Liuzzo
> >> cel (DE): +49 (0) 176 61 000 606
> >> Skype: pietro.liuzzo (Quingentole)
> >> ORCID: https://orcid.org/0000-0001-5714-4011
> >> Academia: https://uni-hamburg.academia.edu/PietroMariaLiuzzo
>
>

Reply via email to