Technical Lab: Manage Virtual Machine Disks
Questions
Question 1 — Multiple Choice
An operations team needs to ensure that data from a critical VM survives accidental deletion of the VM itself. The operating system disk was created along with the VM using Azure portal default settings.
Which statement correctly describes the default behavior and the necessary action to meet the requirement?
A) The managed disk is automatically deleted with the VM by default; it's necessary to uncheck the Delete disk with VM option before deleting the VM.
B) The managed disk is automatically retained after VM deletion; no additional configuration is necessary.
C) Only data disks are deleted with the VM by default; the OS disk is always retained.
D) To retain the disk, it's necessary to convert the managed disk to unmanaged before VM deletion.
Question 2 — Technical Scenario
An administrator receives the following complaint: a production VM with a Premium SSD managed OS disk shows high latency during batch write operations. The VM uses the Standard_D4s_v3 size. After investigation, the administrator discovers that the disk has caching configured as Read/Write.
What is the most likely cause of high latency in this scenario?
A) The Premium SSD type doesn't support batch write operations and should be replaced with Ultra Disk.
B) Read/Write cache on the OS disk accelerates reads, but introduces latency in writes because data needs to be persisted to cache before being confirmed to disk.
C) The Standard_D4s_v3 size is not compatible with Premium SSD disks, causing throttling at the storage layer.
D) Read/Write cache is correct for write workloads; the latency is caused by lack of Write Accelerator configuration on the VM.
Question 3 — True or False
A Ultra Disk managed disk can be resized to increase its capacity without needing to reallocate or recreate the disk, and this operation can be performed with the VM running, without requiring deallocation.
True or False?
Question 4 — Technical Scenario
An administrator needs to move a data disk from a VM located in the East US region to be attached to a VM in the Brazil South region. Both VMs use managed disks. He tries to use the attach existing disk option in the Azure portal, but the destination region doesn't appear as an option.
Source disk: managed disk | East US | Premium SSD LRS
Target VM: Brazil South
What is the cause of the problem and the correct approach to resolve it?
A) LRS managed disks cannot be attached to VMs in other regions; it's necessary to convert to ZRS before moving.
B) Managed disks are regional resources and cannot be directly attached to VMs in another region; it's necessary to create a snapshot of the disk and restore it as a new managed disk in the destination region.
C) The problem is a temporary portal limitation; it's possible to perform the operation via Azure CLI using az disk attach with the --region parameter.
D) It's necessary to export the disk as VHD to a storage account and reimport it in the destination region using the Unmanaged Disk type.
Question 5 — Multiple Choice
An architect compares the available managed disk types to choose the most suitable for a mission-critical OLTP database with requirements of latency below 1 ms and IOPS greater than 80,000.
| Type | Typical latency | Max IOPS (per disk) |
|---|---|---|
| Standard HDD | 10 ms+ | up to 2,000 |
| Standard SSD | 1 to 10 ms | up to 6,000 |
| Premium SSD | < 10 ms | up to 20,000 |
| Ultra Disk | < 1 ms | up to 400,000+ |
Which disk type mandatorily meets both declared requirements?
A) Premium SSD v2, as it offers configurable IOPS above 80,000 with latency below 2 ms.
B) Ultra Disk, as it's the only type that achieves latency below 1 ms and supports IOPS greater than 80,000 per disk.
C) Premium SSD, as with striped disks via Storage Spaces it's possible to achieve the IOPS requirements while maintaining low latency.
D) Standard SSD, as in M-series VMs disk performance is automatically amplified to meet critical OLTP workloads.
Answer Key and Explanations
Answer Key — Question 1
Answer: A
By default, when creating a VM in Azure, the managed OS disk is configured with the Delete disk with VM option enabled. This means that VM deletion results in automatic disk deletion. To retain the disk, the administrator must disable this option before deletion, either in the portal, via CLI (--os-disk-delete-option Detach) or via ARM template.
The main misconception represented by alternatives B and C is assuming that managed disks have automatic protection against deletion, which is not true by default. Alternative D represents an operational setback without benefit; converting to unmanaged disk is not the approach recommended by Microsoft for this scenario.
Answer Key — Question 2
Answer: B
Read/Write cache on the OS disk instructs the host to confirm writes to cache before persisting to physical disk. This improves performance perception for subsequent reads, but in batch write workloads can introduce additional latency because the system waits for write confirmation in cache before proceeding. For OS disks, Read/Write cache is, in fact, the default recommended by Microsoft, but may be inadequate for intensive write workloads.
Alternative C is incorrect because Standard_D4s_v3 supports Premium SSD. Alternative D confuses the role of Write Accelerator, which is a feature available only for M-series VMs and Premium/Ultra disks, not applicable to this context. Alternative A is incorrect as Ultra Disk is not necessarily the solution for this root cause.
Answer Key — Question 3
Answer: False
Although Ultra Disk managed disks support dynamic resizing of capacity, IOPS and throughput, the disk capacity increase operation itself requires the VM to be deallocated at the time of change. This is an important distinction: IOPS and throughput adjustments in Ultra Disk can be made with the VM running, but storage size increase requires deallocation.
Confusing these two behaviors is a common mistake that can cause unplanned unavailability if the administrator tries to resize the disk expecting the operation to occur without downtime.
Answer Key — Question 4
Answer: B
Managed disks in Azure are regional resources. They exist within a specific region and cannot be directly moved or attached to resources in another region. The correct approach is:
- Create a snapshot of the source disk.
- Copy the snapshot to the destination region.
- Create a new managed disk from the snapshot in the destination region.
- Attach the new disk to the destination VM.
Alternative A is incorrect as ZRS geographic replication doesn't solve the attachment limitation between regions. Alternative C is factually wrong: the --region parameter in the az disk attach command doesn't exist for this purpose. Alternative D describes a functional approach in old versions, but reintroduces unmanaged disks, which is contraindicated in current practices.
Answer Key — Question 5
Answer: B
Ultra Disk is the only native Azure managed disk type that simultaneously meets latency below 1 ms and IOPS greater than 80,000 per disk without needing additional striping or aggregation configurations.
Alternative A mentions Premium SSD v2, which indeed offers configurable IOPS and improved latency compared to Premium SSD v1, but its typical latency is approximately 1 to 2 ms, not guaranteeing the strict requirement of below 1 ms. Alternative C transfers complexity to the operating system layer with Storage Spaces, which adds management overhead and doesn't guarantee the required latency. Alternative D is incorrect as M-series with Standard SSD doesn't offer the performance characteristics necessary for this workload profile.