jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jukka Zitting" <jukka.zitt...@gmail.com>
Subject Re: modify a node type definition
Date Thu, 30 Nov 2006 17:40:16 GMT
Hi,

On 11/30/06, Michael Neale <michael.neale@gmail.com> wrote:
> So any hints on how best to design things to make "migration" as easy as
> possible in future? any magic tricks?

Good question. Some rules of thumb:

* When developing your application, use an XML import or a simple
builder application to set up your test content. This way you can
easily scrap the entire test repository, make any modifications (node
type changes, etc.) you need, and then recreate the test repository
without worrying about migrating existing content.

* Avoid putting things directly below the root node. Use a top-level
node like /my:root or /my:content and place all your application
content under that. This way you can easily export/import content
using the standard XML mappings without having to worry about the
protected /jcr:system subtree.

* If your have lots of content, then you should structure it so that
you can export and import it in smaller pieces. If you use references,
make sure that you can import the parts being referenced before the
referencing parts.

* Don't use versioning. As of now Jackrabbit doesn't support migrating
version histories across repositories. This will likely change sooner
or later, but I wouldn't yet count on it.

There is quite a lot of demand for more comprehensive backup/restore
and migration tools so at least I'm quite confident that sooner or
later (hopefully within next year :-) we'll have support for those use
cases. Nicolas' GSoC project was a good start towards achieving that
goal.

BR,

Jukka Zitting

Mime
View raw message