Archiva artifacts on DB or Cloud Storage instead of local file system

[Ralf Mueller <ra...@me.com>]

Hi,

i’d like to implement artifact storage in Archiva on a Database or Cloud Storage (AWS, Google etc.)
instead of the artifacts (.jar, .zip) stored in a directory of the local file system where Archiva is running.

I understand that the metadata can be stored in JCR and/or Cassandra but I’d need a solution where
the artifacts itself should be stored in some DB or Cloud storage. Reason behind is pretty much to support
High Availability in environments where NFS is not available and active-passive not an option.

Wondering if that part of Archiva has some abstractions and where to start looking. I’d definitely contribute
back to Archiva if its not too much of a hassle to implement.

I saw a similar posting about this few months ago but that wasn’t answered very well, hence the question again.

Best Regards,
Ralf

Re: Archiva artifacts on DB or Cloud Storage instead of local file system

[Olivier Lamy <ol...@apache.org>]

Hi,
Well currently there is no abstraction for artifacts storage.
If you want to have a look at the code, the entry point is the class
DefaultRepositoryProxyConnectors (where files are downloaded).
We definitely need some abstractions for file storage.
Feel free to have a look and propose something!

Cheers
Olivier


On 23 September 2016 at 21:28, Ralf Mueller <ra...@me.com> wrote:

> Hi,
>
> i’d like to implement artifact storage in Archiva on a Database or Cloud
> Storage (AWS, Google etc.)
> instead of the artifacts (.jar, .zip) stored in a directory of the local
> file system where Archiva is running.
>
> I understand that the metadata can be stored in JCR and/or Cassandra but
> I’d need a solution where
> the artifacts itself should be stored in some DB or Cloud storage. Reason
> behind is pretty much to support
> High Availability in environments where NFS is not available and
> active-passive not an option.
>
> Wondering if that part of Archiva has some abstractions and where to start
> looking. I’d definitely contribute
> back to Archiva if its not too much of a hassle to implement.
>
> I saw a similar posting about this few months ago but that wasn’t answered
> very well, hence the question again.
>
> Best Regards,
> Ralf
>
>


-- 
Olivier Lamy
http://twitter.com/olamy | http://linkedin.com/in/olamy