s3 with minio

Or: Simple storage for all the things outside git



  • Minio has no history (Amazons own s3 has!)
  • Proper HTTP(S)-based communication protocol
  • Active Community and good documentation
  • Ansibles Modules
  • Standardized API and lots of  tools & clients
  • Single binary deployment (minio)
  • Multinode HA concept build in (minio)
  • Change Monitoring
  • Can be mounted as a file system with minfs
  • May be used for a lot of purposes
  • Easy to use (no git know-how needed)
  • Does not help if files have to be in git repo
  • Official Docker container
  • ONE history for code and binary files
  • Is integrated in git workflow
  • Only one Client
  • Active community and good documentation
  • Integrated in Gitlab
  • Can only be used with git client
  • Not multi purpose, but needed if big files have to be handled inside the repository
  • Git know how needed

(My) conclusion

  • Do we need versions of the same binary file and do we want rollback-features to older versions of this file inside our tower repository?
    • Git LFS!
    • Or... Amazons s3 - because it can do versioning
  • Do we just need an easy to use and well accessible (=http) file storage for all sorts of (big) binary data with high availability features?
    • Let's take s3 by using minio
  • Start single node
  • Start multi node
  • have a look in the cookbooks

Demo Client

  • Show graphical access
    • cyberduck
    • s3explorer
    • web frontend
  • Show mc tool
  • Show curl request
  • Show library access with python
  • Show simple Ansible ad-hoc task

Demo Server