Me too. From my experience the performance of the export to XML is to poor
to be used for a backup.

I don't think JCR2.0 addresses hot backups, so I don't think a standard at
that level will happen. Of course there is nothing stopping jackrabbit
having a persistence manager independent facility (but it would be a LOT of
work, and it has to be near perfect). For now, I would say use the database
facilities (but that also means you can't do a per-workspace hot backup and
restore, in effect).

On 7/25/07, Mark Waschkowski <[EMAIL PROTECTED]> wrote:

I've checked the wiki at:
  http://wiki.apache.org/jackrabbit/BackupTool
and it hasn't been updated since: " 2006-08-16 19:18:27" so I think that
its
way outdated. I looked in Jira and saw that there are some patches, but
its
not complete yet, and nothing added to source tree.

Thomas said:
"Would hot backup of the underlying file system / database be a
solution for you, and if not why not?"
I know for an RDBMS typically you can't do a file system backup of the
database because of issues around transactions not yet committed and the
database state. For postgresql, for example, you can do a hot backup, but
you have to use a database specific tool, not just backing up the files at
any moment in time. Isn't jackrabbit the same in this regard?

I would like to try and pull together some details about this because I'm
sure that there are other people interested in doing hot/backups. If there
is enough input here, I will add or update a backup page into the wiki for
future reference.

Regards,

Mark

Reply via email to