How to make Ceph enterprise ready


Start discussion about steps to make Ceph more enterprise ready and usable for big production deployment e.g. in the telco market. Goal: define subtasks and steps for further development.


  • Danny Al-Gaaf (Deutsche Telekom)
  • Name (Affiliation)
  • Name

Interested Parties

  • Marc Koderer (Deutsche Telekom)
  • Ken Dreyer (Red Hat)
  • Name

Current Status

Detailed Description

Topics to be discuss:
  • The current git repository includes code from several external projects. How can we handle these project version. Should we drop them from source and force the distros to provide these libs as a build dependency to Ceph to increase the maintainability and prevent outdated and no longer supported code in the dev tree? Currently existing projects:
    • libs3 (submodule)
    • civetweb (submodule)
    • jerasure/gf-complete (submodule)
    • rocksdb (submodule)
    • gtest (in-code tree, very bad!)
  • RPM Packages provided via ftp:
    • include also *-devel and src rpms for provided libs like leveldb vs. don't include third party packages and point to the correct 'upstream/distro' repo?
    • how often are these packages synced on ftp from distro-upstream?
  • How can we ensure that a user can always update a production cluster without down time?
  • Feature driven development vs. stability/performance
    • Which features are missing for enterprise use?
    • What tool are missing in the Ceph ecosystem?
  • OpenStack ThirdParty Testing Support for Cinder/Glance/Manila ...
    • Inktank/RedHat should provide a test instance to run devstack against a Ceph cluster
    • provide testcases and report results
  • ... more to be added

Work items

Coding tasks

  1. Task 1
  2. Task 2
  3. Task 3

Build / release tasks

  1. Task 1
  2. Task 2
  3. Task 3

Documentation tasks

  1. Task 1
  2. Task 2
  3. Task 3

Deprecation tasks

  1. Task 1
  2. Task 2
  3. Task 3