The purpose of this event is to introduce interested people to the ceph data-storage ecosystem (1). And have fun ofc ^^
What will happen depends on the people who show up:
Ill give a high level introduction to the components and to why/how ceph works, then everyone can create a ceph cluster with guidance on their notebook(2) or together on a hsspace's test cluster.
We can dive deep into the details:
- How exactly ceph stores data
- What are the advanced usecases and how to adapt the cluster/hw for it
- What are the pitfalls/mistakes that can lead to disaster
- How to debug your cluster
Scenario 3 (reinvent):
With a help of a test cluster we will try to use ceph like noone did it before.
I have a few terrible ideas.
For example: Can you use ceph as an "efficient" data synchronization tool?
Problem: You want to keep a remote copy on a remote host of a directory.
Solution: You run a ceph cluster on your notebook (1 mon/mgr/osd), and you have an osd remotely. You go offline, then you make changes to your filesystem (rbd). The next time you go online, only the changed blocks (objects) will synchronize, giving you an efficient remote copy (as opposed to rsync, but comparable to zfs send).
(1) : In computing, Ceph is a free-software storage platform, implements object storage on a single distributed computer cluster, and provides interfaces for object-, block- and file-level storage. Ceph aims primarily for completely distributed operation without a single point of failure, scalable to the exabyte level, and freely available.
(2) : Requirements: min. 2?gb free ram, gnu/linux!