@stefano, I just watched the recording of Patrick Hausen's @EuroBSDCon distributed storage session and at the end of the session somebody said you might have compared #SeaweedFS with #Garage and that you found SeaweedFS more suitable.

[hnygd@mastodon.africa]

@stefano, I just watched the recording of Patrick Hausen's @EuroBSDCon distributed storage session and at the end of the session somebody said you might have compared #SeaweedFS with #Garage and that you found SeaweedFS more suitable.

I know that you blogged about running SeaweedFS but I wasn't aware you compared different #S3 platforms, if you indeed did, would you mind sharing your experiences?

#FreeBSD

[stefano@mastodon.bsd.cafe]

@hnygd @EuroBSDCon Yes, I talked to Patrick about it shortly after his talk. Basically, I tested both Garage and SeaweedFS in parallel (on the same host) using the BSD Cafe media use case (a huge number of small files). I used rclone to copy data from the Minio instance I was using to both systems. After the first 60 GB, Garage started using all CPU cores and slowed down significantly - to the point that, over an entire night, it copied only a few gigabytes while completely saturating the machine’s CPU resources. SeaweedFS, on the other hand, maintained consistent performance throughout the process. The approach of using "append-only" volumes and small index files is probably better suited for this kind of workload. I was also contacted by some Garage developers who suggested trying the latest versions (which I’ll do as soon as I can), and I’m curious to see if things have improved. I really like Garage and would love to use it more.