Recovery in zonal failure

[VishalBulbule]

Hi Team,

While working on DR design for GCP Compute VM , came across below doubt. It would be great if someone help with this.

VM-1 will be active mode and act as primary in zone-a, Regional disk attached to the primary.

VM-2 will be in standby in zone-b, no extra disk attached.

In case of zona-a failure (network-, plan to detach disk from VM-1 and attach to VM-2 in zone-b.

Here main concern is , if zone-a went down , VM is not accessible , will it allow to detach disk using concole or gcloud command??

[omkarsuram_]

If there is no business reason to have a standby VM, you can spin it up when you detect your primary VM is shutdown. you can save that VM cost, but if this is minimize recovery time as directed by your business goals, then it makes sense. 

When zone goes down you can still access gcp console as its global service. 
For Regional Disks, you can force attach the disk to new VM via console, gcloud or APIs.

When the Primary VM comes back up after zonal recovery, you can use this again to reattach the disk to primary VM.