Data Protection

 View Only

Reduce RTO with Global-Active Device Cloud Quorum

By Kevin Tang posted 07-05-2022 18:00

  

Introduction

What exactly does Global-active device (GAD) cloud quorum do, and who is it for? We will answer both questions by showing you a real-life customer use case where GAD cloud quorum is the perfect solution to address all the customer’s issues with their current environment. After providing an introduction to the standard GAD implementations and the issues that can arise, we’ll show you a side-by-side comparison of the customer’s implementation with and without GAD cloud quorum. If you use, or are interested in GAD, but are limited by a physical quorum disk, read about the progress of GAD quorum options and how the new options make GAD more flexible and resilient.

Background

If you’re unfamiliar with global-active device I suggest taking a look at the following videos:

 

GAD implementations can be challenging for some customers because one of the best practices is using three separate datacenters. This means having your two storage systems and your quorum each at different locations to maximize uptime in the event of site failures such as power outages or natural disasters. Sometimes, we see customers that want to use GAD, but they don't have a third datacenter to host their quorum disk, and we can’t expect them to add a building just to host the quorum. These customers would benefit greatly from being able to locate the quorum on a cloud-based virtual machine.

 

To solve this issue, GAD subject matter expert Dang Luong has been researching and developing procedures since 2019 for configuring various virtual machines as quorum devices. First, an on-premises virtual machine was configured as a quorum, and then the same configuration process was used for AWS and Azure virtual machines. Unfortunately, it takes some time and effort to configure a virtual machine to use as a quorum. So, to make things easier for customers, Jonathan De La Torre and I developed GAD cloud quorum. With GAD cloud quorum, we’ve sped up the configuration process by automating it and packaging it as a cloud marketplace VMI.

 

Here is a timeline of quorum virtual machine offerings and their features.

 

On-premise Fedora Virtual Machine

Amazon Linux VM on AWS

SUSE VM on Azure

GAD Cloud Quorum on AWS/Azure

SUSE VM on Google

Eliminates the need for a third storage system

Eliminates the need for a third datacenter

Eliminates the need for a third datacenter

Eliminates the need for self-service of targetcli configuration

Eliminates the need for a third datacenter

iSCSI connectivity

iSCSI connectivity

 

iSCSI connectivity

 

 

iSCSI connectivity

 

Leverages HA capabilities from hypervisor

 

Leverages HA capabilities from AWS

 

Leverages HA capabilities from Azure

 

Enhances customer experience

 

Leverages HA capabilities from Azure

 

No OS licensing cost

 

No OS licensing cost

 

No OS licensing cost

 

Ensures correct configuration

 

No OS licensing cost

 

Self-service build guide

Self-service build guide (external facing)

Self-service build guide (external facing)

Free on AWS/Azure Marketplace

Self-service build guide (external facing)

 

Before these quorum virtual machines, GAD was still a great solution for some customers with only two data centers. In these scenarios, they could deploy the GAD primary and secondary storage systems at two sites and locate the quorum at one of those sites. Many of these customers have contacted us regarding GAD cloud quorum to increase the resilience of their current GAD configuration.

 

 

You can read more about GAD cloud quorum in co-developer Jonathan De La Torre’s blog post. The main take-away is that customers can easily deploy quorums on AWS and Azure instead of a third data center. In fact, you can go to the marketplace for either cloud provider right now and look up 'Global-active device cloud quorum'. You’ll see a virtual machine image that is ready to serve as a quorum target, and if you want to complete additional configuration, there is a simple command line menu.

 

 

GAD Cloud Quorum Side-by-Side Comparison

As one of the developers for GAD cloud quorum, I was asked to show the customer how migrating their on-premises quorum to the cloud would improve the recovery process. Along with Dang Luong and Tanmoy Panja, I decided to model the customer's configuration in our lab environment.

 

 

 

The procedure included the following two parts:

  1. Simulate an outage of the primary data center (storage system + quorum target) and perform the recovery process.
  2. Migrate to GAD cloud quorum, simulate an outage of the primary data center again, as well as perform the recovery process.

 

For some setups, relocating the quorum to the cloud protects it from a primary data center outage and can be performed in less than an hour!

 

In our tests for this customer’s configuration, migrating to GAD cloud quorum reduced the Recovery Time Objective (RTO) of a data center failure by 50%, and there wasn’t any interruption to the storage system I/O during quorum migration. Overall, this use case served as a great example of how GAD cloud quorum can make GAD more resilient, and we hope that more of our customers will use it to improve their configurations.

 

For more information on GAD cloud quorum on Azure or AWS, see:

GAD Cloud Quorum on Azure

GAD Cloud Quorum on AWS


#GlobalActiveDeviceGAD #ReturnToOriginRTO #SynchronousReplication #DisasterRecovery

​​​​​​​​
11 comments
81 views

Permalink

Comments

11-20-2022 23:42

Excellent stuff; this will help in reducing the RTO.

11-20-2022 22:50

This is a very detailed post on the process

11-20-2022 22:26

Great work on the command line menu for creating quorum volumes.

11-18-2022 13:33

Realizing a 50% RTO reduction by moving to GAD Cloud Quorum shows that businesses can recover from disasters faster and save organizations effort and money spent during the recovery process.  Nicely done, Ken and team.

11-15-2022 14:48

I've often wondered how this worked. Thank you for the update.

11-15-2022 07:28

Good amount of info.

11-15-2022 06:35

Good information about a enhanced feature.

11-15-2022 04:40

Noted, the information!

11-14-2022 22:37

Good amount of info.

11-14-2022 16:04

Great post and guidance. This is applicable to many application deployments such as Oracle RAC. For more information on that, take a look at the following best practices from our UCP for Oracle offering - https://www.hitachivantara.com/en-us/pdf/architecture-guide/protect-database-oracle-12c-with-gad-on-vsp-g-series-best-practices-guide.pdf

11-03-2022 12:10

Nice Post.