Unpacking the "iosxrvk9demo613qcow2 exclusive": What It Is and Why It Matters
QEMU binary: /usr/bin/qemu-system-x86_64
RAM: 4096 MB
Network adapters: 4x virtio-net-pci
Disk image: iosxrvk9demo613qcow2 (exclusive = checked)
Effect: Each instance writes to a copy-on-write overlay, but base image is locked (read-only exclusive). Prevents corruption when multiple labs run concurrently.
“Exclusive” typically means either:
Before importing, ensure your lab environment meets these specs. The 64-bit version is heavier than legacy 32-bit images.
| Setting | Value | Why |
|---------|-------|-----|
| CPU pinning | vcpu.pin0 to physical core 2 | No stealing |
| Hugepages | 1G pages, vm.nr_hugepages=8 | IOS XR memory map fragmentation |
| IOThread | dedicated iothread for qcow2 | QEMU metadata lock avoidance | iosxrvk9demo613qcow2 exclusive
Example exclusive libvirt domain XML snippet:
<cputune>
<vcpupin vcpu='0' cpuset='2'/>
<emulatorpin cpuset='0-1'/>
</cputune>
<memoryBacking>
<hugepages>
<page size='1048576' unit='KiB' nodeset='0'/>
</hugepages>
</memoryBacking>
QCOW2 supports backing files and copy-on-write. If iosxrvk9demo613qcow2 is a base image, you typically create overlay (snapshot) images for each VM instance and lock the base image as exclusive read-only. However, if the base image is accidentally opened read-write by two VMs, corruption is almost certain. Effect: Each instance writes to a copy-on-write overlay,
Thus, hypervisor managers (libvirt, oVirt, OpenStack) use exclusive locks to enforce: