From: Michal Ludvig <michal@lo...> - 2008-06-02 11:22:56
> How is s3fuse moving along?
Hasn't started yet ;-)
> Would some donations help move the project along?
> In particular I would like to see it working under FreeBSD..
In foreseeable future I have a plan to implement S3fs - a full-featured
filesystem using S3 for data storage and possibly Amazon SimpleDB for
metadata (BTW a low-level SimpleDB python class is already in SVN repo).
Full-featured means it will support file renaming, proper directories,
writing into existing files, etc. I wrote up a S3 filesystem
specification some time ago, yet before SimpleDB has been announced.
I'll update it for SimpleDB shortly and publish here for comments.
Once that is done and stable enough I'll connect it with "inotify" (or
similar mechanism on other OSes) to provide online backup. Every file
added or modified in a given directory will automatically get uploaded
into S3. That's close enough to "one way" S3fuse, unless you need to
download data on demand from S3. I.e. in case when the file is not
present in local cache but is available in S3 then S3fuse would download
it for you, whereas S3+inotify won't, it is one way only: local->S3
Only after all that is done I may decide to give S3fuse a go. At the
moment it isn't high on my TODO list though as I believe S3+inotify (or
S3isync as I call it in my mind) is what most people are after anyway.
Would S3isync be enough for what you need Francisco? Or do you need a
real two-ways S3fuse?
Get latest updates about Open Source Projects, Conferences and News.