虚拟机快照和恢复快照
KubeVirt 可以对虚拟机打快照，快照由 snapshot.kubevirt.io 组 API 支持，当前版本为 v1beta1：
[root@base-k8s-master-1 ~]# kubectl api-resources | grep snapshot.kubevirt.io
virtualmachinerestores vmrestore,vmrestores snapshot.kubevirt.io/v1beta1 true VirtualMachineRestore
virtualmachinesnapshotcontents vmsnapshotcontent,vmsnapshotcontents snapshot.kubevirt.io/v1beta1 true VirtualMachineSnapshotContent
virtualmachinesnapshots vmsnapshot,vmsnapshots snapshot.kubevirt.io/v1beta1 true VirtualMachineSnapshot快照依赖 StorageClass，并且 StorageClass 要支持 VolumeSnapshots，还需要配置 VolumeSnapshotClasses。
关于
VolumeSnapshots和VolumeSnapshotClasses的创建可以参考模板，模板地址：https://github.com/ceph/ceph-csi/tree/devel/examples/rbd
[root@base-k8s-master-1 rbd]# kubectl get volumesnapshotclasses.snapshot.storage.k8s.io
NAME DRIVER DELETIONPOLICY AGE
csi-rbdplugin-snapclass rbd.csi.ceph.com Delete 13sKubeVirt 支持在线快照，在线快照依赖
qemu-guest-agent，拍摄快照时虚拟机会冻结，拍摄完成后会解冻。还原快照需要关闭虚拟机后还原。
快照支持包含热插拔磁盘，临时磁盘不支持。
目前快照只能保存磁盘数据，不支持保存内存状态。
虚拟机有多个磁盘时恢复快照后虚拟机可能会从错误的磁盘引导，可以通过
bootOrder设置启动盘。
开启快照功能
需要开启 Snapshot 特性门控。
[root@base-k8s-master-1 rbd]# kubectl edit -n kubevirt kubevirts.kubevirt.io kubevirt
...output omitted...
spec:
certificateRotateStrategy: {}
configuration:
developerConfiguration:
featureGates:
- VMLiveUpdateFeatures
- HonorWaitForFirstConsumer
- Snapshot
...output omitted...为虚拟机打快照前需要先通过
bootOrder字段来设置启动盘，以防恢复后虚拟机从错误的磁盘引导。
为虚拟机打快照
apiVersion: snapshot.kubevirt.io/v1beta1
kind: VirtualMachineSnapshot
metadata:
name: snap-my-vm
spec:
source:
apiGroup: kubevirt.io
kind: VirtualMachine
name: my-vm
failureDeadline: 1m
failureDeadline表示快照超时时间，规定时间内快照未创建完则标记为失败，默认为 5 分钟。
创建快照：
[root@base-k8s-master-1 kubevirt]# kubectl apply -f snapshot.yml
virtualmachinesnapshot.snapshot.kubevirt.io/snap-my-vm created虚拟机删除文件
[root@base-k8s-master-1 kubevirt]# virtctl ssh fedora@my-vm
Last login: Fri Mar 21 16:59:01 2025 from 10.100.239.50
[fedora@vm-dv ~]$ ls /tmp/
systemd-private-e476d875aef945f9a72ebc6e34b57f02-chronyd.service-vJ1UqN systemd-private-e476d875aef945f9a72ebc6e34b57f02-systemd-logind.service-y1erwP
systemd-private-e476d875aef945f9a72ebc6e34b57f02-dbus-broker.service-LNmEGZ systemd-private-e476d875aef945f9a72ebc6e34b57f02-systemd-oomd.service-RO50A9
systemd-private-e476d875aef945f9a72ebc6e34b57f02-systemd-hostnamed.service-VQi9rN systemd-private-e476d875aef945f9a72ebc6e34b57f02-systemd-resolved.service-7ck5Xm
[fedora@vm-dv ~]$ sudo rm -rf /tmp/*
[fedora@vm-dv ~]$ ls /tmp/恢复快照
apiVersion: snapshot.kubevirt.io/v1beta1
kind: VirtualMachineRestore
metadata:
name: restore-my-vm
spec:
target:
apiGroup: kubevirt.io
kind: VirtualMachine
name: my-vm
virtualMachineSnapshotName: snap-my-vm恢复：
[root@base-k8s-master-1 kubevirt]# kubectl apply -f snapshotrestore.yml
The request is invalid: spec.target: VirtualMachine "my-vm" run strategy has to be Halted
[root@base-k8s-master-1 kubevirt]# virtctl stop my-vm
VM my-vm was scheduled to stop
[root@base-k8s-master-1 kubevirt]# kubectl get vm
NAME AGE STATUS READY
my-vm 19d Stopped False
[root@base-k8s-master-1 kubevirt]# kubectl apply -f snapshotrestore.yml
virtualmachinerestore.snapshot.kubevirt.io/restore-my-vm created
[root@base-k8s-master-1 kubevirt]# virtctl start my-vm
VM my-vm was scheduled to start验证
[root@base-k8s-master-1 kubevirt]# virtctl ssh fedora@my-vm
Last login: Fri Mar 21 17:15:10 2025 from 10.100.239.50
[fedora@vm-dv ~]$ ls /tmp/
systemd-private-be271e2ad3734980a387548d0af0351b-chronyd.service-rMAowE systemd-private-be271e2ad3734980a387548d0af0351b-systemd-logind.service-nYHSPP
systemd-private-be271e2ad3734980a387548d0af0351b-dbus-broker.service-8rwdWL systemd-private-be271e2ad3734980a387548d0af0351b-systemd-oomd.service-hejJfq
systemd-private-be271e2ad3734980a387548d0af0351b-systemd-hostnamed.service-lVMJ4U systemd-private-be271e2ad3734980a387548d0af0351b-systemd-resolved.service-1AHHOZTroubleShooting
多个磁盘的虚拟机恢复后未正常引导系统
多个磁盘的虚拟机由于没有设置启动盘，导致恢复快照后虚拟机从数据磁盘引导系统，所以在创建虚拟机时要设置好系统盘，设置系统盘方式如下：
devices:
disks:
- bootOrder: 1
name: root-disk添加 bootOrder 来设置引导顺序。