You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@jackrabbit.apache.org by Lukas Kahwe Smith <sm...@pooteeweet.org> on 2013/06/12 15:18:15 UTC

backup

Aloha,

Currently in production we have been using MySQL for persistence. We now have a client that is already running PostgreSQL. At the same time they told us however that they would actually not mind running with the default XML persistence. However one concern remains backup. It seems like its non trivial to do a hotbackup using XML storage, especially since there will also be binary files (not gigantic files but maybe still small enough to store in PostgreSQL). So it seems like it makes sense to configure Jackrabbit with PostgreSQL mostly to make backup easier.

Correct? Or any recommendations for a hot backup strategy using XML?

regards,
Lukas Kahwe Smith
smith@pooteeweet.org




Re: backup

Posted by Jukka Zitting <ju...@gmail.com>.
Hi,

On Wed, Jun 12, 2013 at 4:18 PM, Lukas Kahwe Smith <sm...@pooteeweet.org> wrote:
> Correct? Or any recommendations for a hot backup strategy using XML?

In general I'd recommend against XML persistence, as it's ancient,
inefficient and as a non-bundle PM deprecated since 2.2 [1]. Use the
embedded Derby PM or BundleFsPersistenceManager if an external
database is not preferred.

On backups, if the underlying storage does not support point-in-time
backups then you'll need an application-level mechanism to guarantee
that no changes are being committed to the repository while a backup
is in progress.

[1] http://archive.apache.org/dist/jackrabbit/2.2.0/RELEASE-NOTES.txt

BR,

Jukka Zitting