A re-introduction to s3fs
summary
s3fs is a FUSE filesystem exposing S3 objects with POSIX semantics but with clear limits: no hard links or atomic renames, costly list/stat because each entry needs a HEAD (multi-rec max default ~20), and random writes amplify to 5MB due to multipart upload rules. It suits sequential reads/writes of bulky files (typical sweet spot 5MB–5GB, supports up to 5TB) and works across many S3 implementations, uses local /tmp caching (flush every ~5GB), recently fixes concurrency/corruption bugs, and often costs far less than NFS-backed cloud storage.
AI-written summary · about
chapters (14)
- 0:00Intro
- 0:37Overview
- 1:47Usage
- 2:23Compatibility
- 4:08Drawbacks
- 5:52Close to Open Data Consistency
- 6:53S3 Implementation Consistency
- 8:16S3fs Performance
- 10:44S3fs Updates
- 12:16S3fs Pitfalls
- 13:19S3fs vs NFS
- 14:17Alternatives
- 15:01Summary
- 15:30Outro
related talks
16mA re-introduction to s3fs
36mFlash Friendly File System
26mImprove FUSE Filesystem Performance and Reliability
19mExtension Framework for File Systems in User space
17mCoFS: A Filesystem for Fast Container Startup
32mBringing up FUSE mounts C/R support
16mSquirrelFS: using the Rust compiler to check file-system crash consistency
30mTo the Cloud and Beyond, Accessing Files Remotely from Linux via SMB3.1.1
