Laurent,
The best way to take a backup is to be a read-transaction and write
n-quads (that's what Fuseki does).
In order to go in at the file level, you need to not have a running JVM.
In 3.1.1, there will be a "lock down" mode to make the on-disk
database consistent, if the app really must.
The best for snapshoting disk files in 3.1.0 is start a write
transaction, do no writes, and take disk copy. This is not a feature
guaranteed in the future.
Taking an RDF-level backup is better.
Andy
On 03/10/16 15:36, Laurent Rucquoy wrote:
Hello,
We have a TDB-backed dataset on a Windows server (Apache Jena 3.1.0)
We want to make a .zip backup of the TDB folder from the JVM running the
TDB.
Is it a safe way to do this in order to avoid database corruption ?
Thank you in advance for your help.
Regards,
Laurent
