AWS S3 vs ...?

[ljohnston]

Hi,

I recently came across pupistry and have been playing with it a little bit. I've only scratched the surface, but it seems like a really nice approach to masterless puppet. Something I'm wondering about... did you give any thought to something other than S3 for storing artifacts (e.g. an http server or binary artifact repository)? Currently S3 isn't a problem for us, but having yet one more thing tying us to AWS is less than ideal. I'd also suspect that for some, this may be a barrier to entry.

Any possibility pupistry could support multiple backends for artifact storage?

Thanks for an intriguing tool that I'm looking forward to learning more about!

Regards,
Lance

[jethrocarr]

hi Lance,

There's no specific reason it has to be S3 - I picked S3 since it's probably the most common object store currently used and made it easy to get started, but I'd be happy to accept pull requests adding support for alternative platforms.

It wouldn't be too hard, we would need to introduce a generic storage class to be called (in place of StorageAWS) which can then alias to whatever storage backend is select, eg StorageAWS, StorageHTTP, StorageHyperCloud9000. I kept things mostly isolated, I just never bothered writing the generic layer yet.

I don't have need for it so won't be writing it myself any time soon, but I'll happily advise on how it could be done and review/merge any PRs that come my way.

regards,
jethro

[ljohnston]

Hey Jethro,

Thanks for the response. Like I said, S3 isn't currently an issue for us, so I'm pretty much in the same boat you are at the moment... don't need it so probably won't be implementing anytime soon. If I get there, however, I'll definitely take you up on your offer for advise.

Thanks,
Lance