How delete binary nodes immediately also on filesystem?

[Tino Rink <ti...@gmail.com>]

Hallo Everyone

We would like to use Jackrabbit in our current project. One of our
requirements is to store many documents and some with private or secure
nature. If a user request the deletion of such a sensitive document, it
should happend immediately - even from the filesystem.

After reading the documentation (
http://wiki.apache.org/jackrabbit/DataStore#Data_Store_Garbage_Collection)
it seem there only one way:
1. deleting the node including binary data property
2. run the GC

Questons:
- Is there another / direkt way to delete the binary content from fs?
- How 'fast' is the GC when talking about GB till TB of Data?
- Could the GC-Run be resticted / advised to analyse and process only a
subset of nodes (maybe only one)?


Thanks and best regards,
Tino

Re: How delete binary nodes immediately also on filesystem?

[Martin Bonato <m....@datentechnik-innovation.com>]

Hi Tino!

I don't know if the GC could be used in such a way, but considering your requirements I would suggest you encrypt the documents with a generated session key and store the key as a property of your documents nodes. Now you only need to overwrite/delete the session key and the (encrypted) document is picked up by the GC later. Without the key however, the document can not be recovered.

HTH,
Martin

On 13.10.2014, at 16:41, Tino Rink <ti...@gmail.com> wrote:

> Hallo Everyone
> 
> We would like to use Jackrabbit in our current project. One of our
> requirements is to store many documents and some with private or secure
> nature. If a user request the deletion of such a sensitive document, it
> should happend immediately - even from the filesystem.
> 
> After reading the documentation (
> http://wiki.apache.org/jackrabbit/DataStore#Data_Store_Garbage_Collection)
> it seem there only one way:
> 1. deleting the node including binary data property
> 2. run the GC
> 
> Questons:
> - Is there another / direkt way to delete the binary content from fs?
> - How 'fast' is the GC when talking about GB till TB of Data?
> - Could the GC-Run be resticted / advised to analyse and process only a
> subset of nodes (maybe only one)?
> 
> 
> Thanks and best regards,
> Tino