CPU Type Changes: Pass-through native to VM Guests

Ended
Resolved

The changes & power cycles have been successfully completed. Thanks for your patience!

Began
Maintenance Planned

G’day,

We’ll be performing a small change to every Hosted Guest (VM / VPS / etc) that we operate. This will involve changing the presented silicon from being a QEMU/KVM type to passing through the host node’s real CPU type.

The key benefits here are getting as many functions and instruction sets enabled natively, without needing to manually configure this per-guest on top of the QEMU default emulation type. For instance, we already have security mitigation flags enabled and routed into guests, however this change will cover that by default as well as other sets like AVX - which is starting to become mandatory for system upgrades like for MongoDB, CL, etc.

For each guest, we will amend the hypervisor config for it, then shut-down guest, then power-on guest. Done!

If you’ve any questions about these works, please let us know. Our goal is to complete it early in the weekend.

Cheers,
Merlot Digital

8 Affected Services:
The Network Crew Pty Ltd (TNC)

« Merlot Digital website

Network: AS138521