Ceph RBD for KVM and libvirt virtual machines
Running VM disks on RBD without a hypervisor platform: libvirt secrets for cephx users, QEMU librbd disks, cache modes, live migration between hosts and pool layout.
Ceph block storage (RBD)
Specific tasks for Ceph RBD block storage: virtual machine disks, noisy neighbors, client caching, snapshots and replacing the iSCSI gateway.
Ceph RBD QoS limits for noisy neighbors
Capping IOPS and throughput per image: combined and per-direction limits, bursts, where limits apply in librbd, and why kernel RBD clients need other controls.
Replacing the Ceph iSCSI gateway
The iSCSI gateway is in maintenance mode: moving VMware hosts to the NVMe-oF gateway, Linux hosts to native RBD, and planning the cutover without downtime.
librbd cache settings: writeback, writearound and writethrough
Choosing a client cache policy for VM workloads: defaults, cache sizes and dirty limits, writethrough until flush, and the data safety trade-offs.
Ceph RBD snapshots and clones for VM templates
Using snapshots for golden images and fast clones: protected snapshots, flattening, the effect of deep clone chains and using snapshots as backup sources.
