host passthrough not working with KVMs
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
MAAS |
Fix Committed
|
Medium
|
Unassigned |
Bug Description
When deploying a KVM VM, the CPU capabilities are very limited and set to QEMU/KVM CPU. I’d like to use host-passthrough as a default but I’m not sure where to find how MAAS communicates with Virsh or any config files.
I’d like the virtual machines that MAAS creates to have host-passthrough as the CPU arch instead of the current CPU arch which is very basic.
For example, creating a VM with MAAS will produce a virtual machine that can’t even run Docker because the VCPU doesn’t have SSE4 capabilities.
<cpu mode='host-
I’d like to be automatically inserted into each new domain that is created by MAAS
Just to confirm, we’re not talking about the CPU architecture as in x86_64 but the feature flags and processor name/details retrieved from /proc/cpuinfo. I’d like them to have the host flags and details as via host-passthrough.
Related branches
- Anton Troyanov: Approve
- MAAS Lander: Approve
-
Diff: 228 lines (+78/-50)2 files modifiedsrc/provisioningserver/drivers/pod/virsh.py (+52/-50)
src/tests/provisioningserver/drivers/pod/test_virsh.py (+26/-0)
tags: | added: bug-council |
Changed in maas: | |
status: | Expired → Triaged |
Changed in maas: | |
status: | Triaged → Fix Committed |
Changed in maas: | |
milestone: | 3.5.0 → 3.4.0 |
status: | Fix Committed → Fix Released |
status: | Fix Released → Fix Committed |
Changed in maas: | |
milestone: | 3.4.0 → 3.5.0 |
Changed in maas: | |
milestone: | 3.5.0 → 3.4.0 |
Changed in maas: | |
milestone: | 3.4.0 → 3.4.0-beta2 |
How does this work for LXD VM hosts?