Deep Dive into Ceph Block Storage
summary
Explains how RBD stores thin-provisioned images as objects mapped across placement groups and OSDs, enabling striping across OSDs (default 4 MiB object/stripe size, stripe_count=1) for RAID0-like throughput; details snapshot/clone semantics, copy-on-write behavior, and the need to protect/unprotect and flatten clones with deep flatten keeps old snapshots valid; compares librbd vs kernel rbd, outlines rbd-mirror replication and experimental caches, and lists exact libvirt/qemu steps and secret setup to attach images to VMs.
AI-written summary · about
chapters (8)
- 0:00Introduction to Ceph
- 1:01RBD images and flow
- 2:51Core RBD features
- 4:33Image formats and mirroring
- 6:12RBD image components
- 7:22Striping and snapshots
- 10:32Cloning and synchronization
- 15:40Virtualization integration
related talks
19mMicroCeph: Simplifying Storage
10mBring SCSI Support Into QEMU Block Layer
48mSoftware Defined Storage the Linux way
30mQEMU-Storage-Daemon and libblkio: Exploring New Shores for the QE...
18mDm-qcow2: qcow2 as Device-mapper Block Device
21mKeynote: Ceph Developers AMA: Meet the Minds Behind Open Source Ceph
52mPeeling Back the Layers of Storage
31mEROFS and containers