Incremental snapshots/backups in KVM plugin of Apache CloudStack
summary
Implements native incremental backups for KVM in Apache CloudStack by using libvirt’s begin_backup API with QCOW2 backing chains so only changed blocks are copied; requires libvirt 7.6 and QEMU 6.1 and currently supports file‑based stores (local, NFS, RBD). Demonstrates a concrete savings example: a 500 GB volume growing 20 GB/day uses ~2 TB with full daily snapshots vs ~560 GB with incrementals (≈25%). Uses dummy VMs for detached volumes and enforces a chain limit 16 to avoid long merge times.
AI-written summary · about
chapters (7)
- 0:00Introduction to snapshots
- 3:03Incremental snapshots concept
- 6:00Snapshot chain management
- 7:26Creation workflow details
- 8:52Deleting and reverting snapshots
- 10:38Feature limitations and requirements
- 12:48Q&A and clarifications
related talks
20mFaster Backups, Less Storage: Unlocking Incremental Backups in KubeVirt
27mBitmaps and NBD: Building Blocks of Change Block Tracking
20mFrom Agent-Based to Agentless...
16mProxmox Backup Server: Backup for your Datacenter
28mCSI Volume Snapshots: On the Way to Faster and Better Backups
27mIncremental backups in Kubernetes: Where we stand and what's missing
49mIncremental Backup
26mLessons Learned: Optimizing KVM Performance for EHR Systems