Persistent disk redundancy across multiple regions

10/14/2019

Is there a way to create a Persistent Disk (disk_a) in one GCP region (us-central1) and maintain a copy(disk_b) in another GCP region(asia-south1)?

Is it possible using some Distributed file system like Ceph or GlusterFS or anything else?

This is required so that in a rare case of regional failure (I understand it is barely possible), I can create a cluster or VM in the other region and mount the other disk which has the same data?

-- Amit Yadav
ceph
glusterfs
google-cloud-platform
google-kubernetes-engine
persistent-storage

1 Answer

10/14/2019

In Google Cloud Platform, there are global, regional and zonal resources. Persistent disks are regional or zonal. Global (multi-region) is not offered at this time.

For the booting disk drive, you must select from regional or zonal persistent disks or local storage. Global is not an option. You can implement global file systems using your own custom services, but you cannot boot a compute engine instance from them. You can only attach additional storage from custom services.

For disaster recovery, you can create snapshots. These are global resources. How often to create snapshots depends on your RTO and RPO, the shorter these goals the more it will cost you in both money and management.

There are much better technologies to use for HA and DR than globally persistent disks that are more cost-effective and implementation efficient.

-- John Hanley
Source: StackOverflow