How is Swift handling files with the same content/hash

asked 2018-06-10 04:23:15 -0600

I am using the OVH Object Storage. I uploaded quite a lot of files. A high percentage of these files have the same content/md5 hash, but a different name.

I checked the size of the container compared to the files on my disk. There seems to be a big difference in file size. Therefore my assumption would be that if uploading multiple files with the same hash, Swift only puts the pointer to the same file instead of writing multiple files with the same content. This wouldn't completely work with the hash ring, but I uploaded many files (>10.000.000) so the likelihood of sharing the same file per cluster in the ring seems still kind of likely.

Is this true or is there a different reason for the difference in file size?

edit retag flag offensive close merge delete