Client Driven Geo-Redundancy for Ceph Object Storage
summary
A Ceph RGW extension adds a new object_synced_create bucket-notification event that is emitted from the destination zone (configured on Zone B) when asynchronous multisite replication finishes, and it delivers JSON to a Kafka topic for clients to consume. Demo runs two RGWs on one host (ports 8101 and 8201) and shows the consumer receiving the sync event after upload. A draft Boto3 wrapper and proposals for metadata-sync and deletion events exist; demo proves the mechanism works.
AI-written summary · about
chapters (6)
- 0:00Introduction and motivation
- 1:51Client control over replication
- 3:58How notifications work
- 5:43Live demonstration setup
- 8:12Future improvements and next steps
- 11:06Q&A session
related talks
35mKeeping Ceph RGW Object Storage Consistent
25mEvolution of Geo-replication in Gluster
20mRemote Replication in MicroCeph: RBD and Beyond
25mSupporting 3 Availability Zones Stretch Cluster
02h 59mCeph Storage: a Caffeinated Primer
34mModernizing Ceph Deployments at CERN: CephFS and Object Storage Across Data Centres
44mEasy Geo-Redundancy with MARS
29mGlobal Deduplication for Ceph
