Using Subsonic with Amazon S3 for Storage - Issues

[andermic]

I did a lot of research before I tried this out, and I have it kind of working, but I still have some issues. I have already read through this post: http://forum.subsonic.org/forum/viewtopic.php?t=5411

This is how I have it set up. I have the basic VPS from linode running with subsonic, and I have amazon s3 mounted with s3fs. It's all pretty slick, and I really like the layout, but I'm having one issue that is keeping me from fully implementing this.

The major issue that I'm having is that the combination of Subsonic and S3FS seems to pull all of the data at once, for some reason. I only have about 2 GBs of music on the server right now, just to try it out. And it seemed that when I loaded subsonic, it pulled over 30GB of traffic from my Amazon S3 instance. I have no idea what it would have to do to get that much data. Also, I got alerts from linode that I was causing a heavy load on my server. I enabled the local caching option for s3fs and it has fixed the problem, however it instantly pulled down all 2GB of music and stopped. This completely defeats the purpose of having the data stored on Amazon. So somewhere, subsonic is doing something that is requiring s3fs to pull all of the data.

Does anyone know what might be causing this? And if there would be a solution for it? I would really like to keep this setup, but I might have to switch to something else. I like this because then my music is backed up, and I don't have to keep a computer on at home (I only have a laptop right now).

Thanks.

[zapt0]

It's most likely because Subsonic reads all the tag data (for which it needs access to the entire file) during library refreshing.

After the data has been pulled once (2GB total in your case) it should not have to be pulled again. However, if the file timestamps of your mounted file system change, it might trigger Subsonic to redownload the file.