An organization hosts a critical financial web application in the Canada Central region. T… — Microsoft Azure Solutions Architect (AZ-305)

An organization hosts a critical financial web application in the Canada Central region. The application architecture consists of the following components:

- Two web server virtual machines (`Web-VM1` and `Web-VM2`), each containing a single $127\text{ GB}$ Standard SSD disk with a write churn of $1.5\text{ MB/s}$ per disk.
- One database virtual machine (`DB-VM1`) running Microsoft SQL Server with a total VM write churn of $50\text{ MB/s}$ . The virtual machine has three disks: an OS disk (Standard SSD, $2\text{ MB/s}$ write churn), a database data disk (Premium SSD v1, $14\text{ MB/s}$ write churn), and a transaction log disk (Premium SSD v1, $34\text{ MB/s}$ write churn).

The organization is designing a disaster recovery (DR) solution to the Canada East region. The design must meet the following objectives:

- A recovery time objective (RTO) of $2\text{ hours}$ for the entire application.
- A recovery point objective (RPO) of $4\text{ hours}$ for the web servers.
- A recovery point objective (RPO) of $10\text{ seconds}$ for the database server.

Which disaster recovery strategy should you recommend?

Configure Azure Site Recovery to replicate the web server virtual machines to the Canada East region. Deploy a secondary SQL Server virtual machine in the Canada East region and configure asynchronous SQL Server Always On Availability Groups to replicate the database.Answer
B
Configure Azure Site Recovery to replicate all virtual machines, including the web servers and the database server, using a replication policy with multi-VM consistency enabled for the database tier.
C
Configure Azure Site Recovery to replicate the web server virtual machines. Configure Azure Backup for the database virtual machine with Cross-Region Restore enabled to replicate database recovery points to the Canada East region.
D
Configure Azure Site Recovery to replicate the web server virtual machines. Store the database files on a Read-Access Geo-Redundant Storage account and mount the storage to a recovery virtual machine in the Canada East region.

Answer

Configure Azure Site Recovery to replicate the web server virtual machines to the Canada East region. Deploy a secondary SQL Server virtual machine in the Canada East region and configure asynchronous SQL Server Always On Availability Groups to replicate the database.

The correct strategy replicates the web servers using Azure Site Recovery and utilizes SQL Server native replication (Always On Availability Groups) for the database. This is necessary because the transaction log disk's write churn of

34\text{ MB/s}

exceeds the limits of Azure Site Recovery (which tops out at

30\text{ MB/s}

per Premium SSD v1 disk even under the high-churn policy). Furthermore, Azure Site Recovery generates crash-consistent recovery points at a minimum interval of

5\text{ minutes}

, which cannot satisfy the strict

10\text{ second}

RPO requirement. Deploying native database replication resolves both the churn threshold constraint and the low RPO objective.

Step-by-Step Solution

Analyze the database disk write churn and total VM write churn against Azure Site Recovery limits.

The database transaction log disk write churn of

34\text{ MB/s}

exceeds the maximum supported single Premium SSD v1 disk limit (

20\text{ MB/s}

for standard tier,

30\text{ MB/s}

for high-churn tier).

This determines if Azure Site Recovery is a viable replication option for the database virtual machine.

Compare the RPO requirements against the replication capabilities of Azure Site Recovery and Azure Backup.

The database RPO of

10\text{ seconds}

cannot be met by Azure Site Recovery (which has a recovery point frequency of several minutes) or Azure Backup (which operates on hourly/daily snapshots).

This rules out Azure Backup and standard virtual machine-level replication for the database tier.

Evaluate native database replication options for the SQL Server workload.

SQL Server Always On Availability Groups configured in asynchronous commit mode can continuously replicate database transactions to another region, satisfying the

10\text{ second}

RPO.

Native database replication is required to achieve low RPOs and handle high write churn workloads.

Select the correct combination of replication tools for the multi-tier application.

Use Azure Site Recovery for the low-churn web virtual machines (which easily meet the

4\text{ hour}

RPO) and native SQL Server replication for the database tier.

This architecture meets both the workload performance limitations and the RPO/RTO business requirements.

Key Concept

Azure Site Recovery capacity limits and native database replication for low recovery point objectives.

Question

Answer

Step-by-Step Solution

Key Concept

Common Mistakes

Topics

More questions from this topic