Pass Your Microsoft AZ-104 Exam Easy!

Changing VM size is a disruptive compute operation, even when the requested size is available without deallocation; a restart may occur. The question also states the desired size is unavailable on the existing hardware cluster, so a deallocation or suitable supported relocation may be necessary. Performing an unscheduled resize while assuming no service interruption creates avoidable operational risk and does not solve the placement restriction. Find a supported SKU and plan controlled deallocation, capacity checks, restart and application-health verification before requesting the final size.

Regional SKU availability alone does not imply that the currently allocated host cluster supports a requested resize. Azure may require deallocation to release the VM and allocate its new size on another supported cluster, and existing availability-set constraints can still matter. Here the specific size is already reported unavailable on the present cluster. Identify sizes supported for this VM and the applicable move or deallocation path, including downtime and any placement constraints, instead of equating region-wide catalog presence with immediate in-place resize capacity.

Storage performance tier controls disk throughput and IOPS characteristics; it does not change the number of virtual CPUs allocated to a VM. The workload needs more vCPUs, and its target compute size is unavailable on the current cluster. Increasing disk performance may be a useful remedy for an I/O-bound workload but cannot solve a CPU-capacity or cluster-availability constraint. Evaluate supported VM sizes and any required deallocation or reallocation procedure, then verify CPU resources and application health after the change.

'Check OS, architecture, generation and disk/SKU compatibility for the target VM size' applies to a different feature boundary because an available SKU is not automatically compatible with the VM's image and boot requirements. In the stated scenario where a VM size change is blocked with a vCPU quota error in a particular VM family, 'Check OS, architecture, generation and disk/SKU compatibility for the target VM size' does not satisfy the mechanism that quota can block deployment even when the requested size exists and RBAC permits management. The necessary action is 'Review regional VM-family vCPU quota and request an approved increase', which should be checked against this Azure resource's actual permissions and state.

The nearby action 'Choose a compatible memory-optimized VM size rather than only increasing disk or CPU' is associated with this separate need: an analytics application repeatedly exhausts RAM but has spare CPU. That is because memory-bound demand calls for more RAM within regional quota and cost constraints. Here a VM size change is blocked with a vCPU quota error in a particular VM family, so 'Choose a compatible memory-optimized VM size rather than only increasing disk or CPU' misses the fact that quota can block deployment even when the requested size exists and RBAC permits management. 'Review regional VM-family vCPU quota and request an approved increase' is the more defensible action; confirm its behavior in the target service before calling the task complete.

'Choose a steady-performance VM family rather than relying on limited burst credits' solves another task because burstable credits may be depleted during continuous load, reducing sustained performance. For the situation where a VM size change is blocked with a vCPU quota error in a particular VM family, this distractor misses the requirement that quota can block deployment even when the requested size exists and RBAC permits management. Unlike 'Choose a steady-performance VM family rather than relying on limited burst credits', 'Review regional VM-family vCPU quota and request an approved increase' directly targets that requirement; its success must be confirmed at the authorized resource scope.

A supported size on the current hardware cluster can avoid the separate need to deallocate and reallocate the VM, but it does not make resizing a running Azure VM nondisruptive. Microsoft documents that a running VM restarts during the size-change operation. An application may lose active sessions or need recovery time even when the SKU and quota were validated successfully. This answer incorrectly equates avoiding deallocation with avoiding an operating-system restart, contrary to the explicit requirement to plan a maintenance interruption.

VM sizing changes the virtual machine's allocated compute resources, such as vCPU and memory, rather than acting as a data-disk capacity expansion. It is not a guest filesystem resize operation and does not promise that the operating system continues running without interruption. For a running Azure VM, Microsoft specifies a restart during resize. Any disk or filesystem growth would be a different planned operation with its own guest steps. The application team should therefore expect interruption and perform post-resize service validation.

Deallocation can be required when the desired size is unavailable on the physical cluster hosting the current VM. However, Microsoft explicitly allows a running VM to be resized to a supported available size without a separate preliminary deallocation step; that resize still restarts it. This scenario removes the cluster-availability uncertainty by stating that the target size was verified there. Saying all resizes require deallocation makes a conditional operational prerequisite universal, rather than identifying the unavoidable restart and application validation that the maintenance plan needs.

Generation 1 and Generation 2 VMs use different firmware and boot characteristics; changing a boot disk's generation is not a routine in-place resize operation. The question states the existing VM is Generation 2 with Trusted Launch, and the candidate family lacks the required security support. Converting boot generation without a planned image or redeployment procedure can make the workload unbootable and still does not establish support for the desired target SKU. The administrator should first select a size compatible with the VM's existing generation, security type, operating system and region rather than modifying the disk to fit an unsupported family.

Regional and family vCPU quotas govern how much compute capacity a subscription may allocate, but they do not determine whether a VM size implements Trusted Launch, Secure Boot or vTPM. A quota increase might address a separate resource allocation denial when the SKU is otherwise supported. In this case the family is available in the region but is explicitly unsupported for Trusted Launch, which is a feature compatibility problem. Check the VM's effective security configuration and a supported size family before requesting capacity. Do not represent a quota approval as a way to enable missing hardware or firmware security capabilities.

Trusted Launch security features may be business or compliance requirements. Disabling Secure Boot and virtual TPM by default across a production fleet could reduce security protections and is not an appropriate response to one proposed incompatible VM resize. Microsoft documents specific workflows and restrictions for changing a VM's security type; support depends on its image, generation, client tooling and configuration. The problem does not authorize a downgrade or show that doing so is necessary. Prefer a supported VM size preserving existing Trusted Launch controls and record any exception as an independently reviewed security decision.

'Check disk encryption, region, state and attachment compatibility before migration' solves another task because disk availability and key access matter; a new VM cannot always attach arbitrary encrypted disks. For the situation where a database VM's data volume is full while its operating system disk has space, this distractor misses the requirement that OS and data disks hold different filesystems; adding OS capacity may not fix database storage. Unlike 'Check disk encryption, region, state and attachment compatibility before migration', 'Expand the correct managed data disk, then extend the guest volume if required' directly targets that requirement; its success must be confirmed at the authorized resource scope.

When a team needs to restore an entire corrupted VM with policy retention, rather than one disk point-in-time copy, the action 'Use Azure VM Backup rather than relying only on an isolated managed-disk snapshot' can be relevant: VM backup manages recovery points, schedule and restore workflow; disk snapshots are not equivalent coverage. For the different problem of a database VM's data volume is full while its operating system disk has space, 'Use Azure VM Backup rather than relying only on an isolated managed-disk snapshot' misses the mechanism that OS and data disks hold different filesystems; adding OS capacity may not fix database storage. Choosing 'Use Azure VM Backup rather than relying only on an isolated managed-disk snapshot' would address that other need, whereas 'Expand the correct managed data disk, then extend the guest volume if required' is intended to solve the stated condition.

The database’s data volume is on a distinct managed data disk. Increasing the OS disk does not enlarge the full data filesystem that receives database writes. Select the actual managed data disk, verify its supported limits and resize it, then extend the relevant partition or filesystem in the guest as required. Capacity on a different disk cannot satisfy this storage-pressure condition.

'Evaluate Azure VM encryption at host on supported configurations' solves another task because host encryption protects host-managed data paths; Azure Disk Encryption works through guest OS mechanisms. For the situation where a database VM's data volume is full while its operating system disk has space, this distractor misses the requirement that OS and data disks hold different filesystems; adding OS capacity may not fix database storage. Unlike 'Evaluate Azure VM encryption at host on supported configurations', 'Expand the correct managed data disk, then extend the guest volume if required' directly targets that requirement; its success must be confirmed at the authorized resource scope.

When a team needs to restore an entire corrupted VM with policy retention, rather than one disk point-in-time copy, the action 'Use Azure VM Backup rather than relying only on an isolated managed-disk snapshot' can be relevant: VM backup manages recovery points, schedule and restore workflow; disk snapshots are not equivalent coverage. For the different problem of a transactional VM is IOPS-limited despite low CPU usage, 'Use Azure VM Backup rather than relying only on an isolated managed-disk snapshot' misses the mechanism that disk performance and VM limits both constrain I/O; more vCPUs alone may not help. Choosing 'Use Azure VM Backup rather than relying only on an isolated managed-disk snapshot' would address that other need, whereas 'Select a supported higher-performance disk tier and confirm VM throughput limits' is intended to solve the stated condition.

The nearby action 'Choose a steady-performance VM family rather than relying on limited burst credits' is associated with this separate need: a production service has sustained high CPU all day; a burstable B-series VM is being proposed solely for low price. That is because burstable credits may be depleted during continuous load, reducing sustained performance. Here a transactional VM is IOPS-limited despite low CPU usage, so 'Choose a steady-performance VM family rather than relying on limited burst credits' misses the fact that disk performance and VM limits both constrain I/O; more vCPUs alone may not help. 'Select a supported higher-performance disk tier and confirm VM throughput limits' is the more defensible action; confirm its behavior in the target service before calling the task complete.

'Check disk encryption, region, state and attachment compatibility before migration' solves another task because disk availability and key access matter; a new VM cannot always attach arbitrary encrypted disks. For the situation where a transactional VM is IOPS-limited despite low CPU usage, this distractor misses the requirement that disk performance and VM limits both constrain I/O; more vCPUs alone may not help. Unlike 'Check disk encryption, region, state and attachment compatibility before migration', 'Select a supported higher-performance disk tier and confirm VM throughput limits' directly targets that requirement; its success must be confirmed at the authorized resource scope.

'Verify guest device discovery, mount point, filesystem and permissions' applies to a different feature boundary because ARM attachment is only the infrastructure step; the guest must recognize and mount the volume correctly. In the stated scenario where an admin enlarged a managed disk in Azure, but the Windows guest still shows the old filesystem size, 'Verify guest device discovery, mount point, filesystem and permissions' does not satisfy the mechanism that cloud disk expansion does not always automatically grow the guest partition or filesystem. The necessary action is 'Extend the guest partition and filesystem into the new capacity', which should be checked against this Azure resource's actual permissions and state.

Azure-side disk enlargement increases the block device capacity, but the Windows guest may still expose the old filesystem size until its partition and volume are extended. Repeating the cloud disk resize without updating the guest leaves the observable symptom unresolved. Inspect disk/volume state inside the VM and extend the appropriate partition into the already provisioned free space before retesting application capacity.

When an encrypted managed disk is moved for attachment to a different VM, the action 'Check disk encryption, region, state and attachment compatibility before migration' can be relevant: disk availability and key access matter; a new VM cannot always attach arbitrary encrypted disks. For the different problem of an admin enlarged a managed disk in Azure, but the Windows guest still shows the old filesystem size, 'Check disk encryption, region, state and attachment compatibility before migration' misses the mechanism that cloud disk expansion does not always automatically grow the guest partition or filesystem. Choosing 'Check disk encryption, region, state and attachment compatibility before migration' would address that other need, whereas 'Extend the guest partition and filesystem into the new capacity' is intended to solve the stated condition.

The nearby action 'Use additional VM instances for horizontal scale-out, not a larger single disk' is associated with this separate need: a VM disk is enlarged and the team calls that scale-out. That is because horizontal scaling adds compute replicas; disk expansion is storage capacity vertical change. Here a team needs to restore an entire corrupted VM with policy retention, rather than one disk point-in-time copy, so 'Use additional VM instances for horizontal scale-out, not a larger single disk' misses the fact that VM backup manages recovery points, schedule and restore workflow; disk snapshots are not equivalent coverage. 'Use Azure VM Backup rather than relying only on an isolated managed-disk snapshot' is the more defensible action; confirm its behavior in the target service before calling the task complete.

'Check disk encryption, region, state and attachment compatibility before migration' solves another task because disk availability and key access matter; a new VM cannot always attach arbitrary encrypted disks. For the situation where a team needs to restore an entire corrupted VM with policy retention, rather than one disk point-in-time copy, this distractor misses the requirement that VM backup manages recovery points, schedule and restore workflow; disk snapshots are not equivalent coverage. Unlike 'Check disk encryption, region, state and attachment compatibility before migration', 'Use Azure VM Backup rather than relying only on an isolated managed-disk snapshot' directly targets that requirement; its success must be confirmed at the authorized resource scope.

When a database VM's data volume is full while its operating system disk has space, the action 'Expand the correct managed data disk, then extend the guest volume if required' can be relevant: OS and data disks hold different filesystems; adding OS capacity may not fix database storage. For the different problem of a team needs to restore an entire corrupted VM with policy retention, rather than one disk point-in-time copy, 'Expand the correct managed data disk, then extend the guest volume if required' misses the mechanism that VM backup manages recovery points, schedule and restore workflow; disk snapshots are not equivalent coverage. Choosing 'Expand the correct managed data disk, then extend the guest volume if required' would address that other need, whereas 'Use Azure VM Backup rather than relying only on an isolated managed-disk snapshot' is intended to solve the stated condition.

'Select a supported higher-performance disk tier and confirm VM throughput limits' applies to a different feature boundary because disk performance and VM limits both constrain I/O; more vCPUs alone may not help. In the stated scenario where a team needs to restore an entire corrupted VM with policy retention, rather than one disk point-in-time copy, 'Select a supported higher-performance disk tier and confirm VM throughput limits' does not satisfy the mechanism that VM backup manages recovery points, schedule and restore workflow; disk snapshots are not equivalent coverage. The necessary action is 'Use Azure VM Backup rather than relying only on an isolated managed-disk snapshot', which should be checked against this Azure resource's actual permissions and state.

Encryption at host covers supported VM-host paths such as OS/data disk caches and temporary disks. It is not a replacement for the customer-managed key protecting an already encrypted managed data disk through a disk encryption set. Turning on host encryption for a target VM cannot repair a disabled Key Vault key or revoke a DES identity dependency. The administrator must verify that the destination VM can attach the disk and that Azure Disk Storage can still access the DES and its key. Host-encryption prerequisites, such as VM family support and prior encryption history, should be evaluated independently if that feature is also required.

Copying data to a different disk would be a separate migration requiring a secure, consistent data-transfer and recovery plan. It is not a supported shortcut for silently stripping the existing customer-managed disk encryption dependency in the scenario. The managed disk remains protected by a DES whose managed identity must perform authorized cryptographic operations against the configured key. Removing or losing those permissions can render encrypted data inaccessible. Keep the required key service available, validate disk/VM attachment compatibility and use a formally approved migration procedure if encryption architecture truly must change.

Azure Reader at the subscription scope allows management-plane visibility to permitted resources; it does not grant the cryptographic operations that the disk encryption set identity requires for a customer-managed encrypted disk. Even a wider management role would not magically re-enable a disabled key or repair incompatible attachment location and disk state. For the specified cross-VM attachment, the operator must verify the disk encryption set, managed identity Key Vault access and supported disk/VM placement first. Give the human operator only the necessary administrative rights for the attachment process, separate from the encryption service's key access.

'Expand the correct managed data disk, then extend the guest volume if required' applies to a different feature boundary because OS and data disks hold different filesystems; adding OS capacity may not fix database storage. In the stated scenario where a disk appears attached to a VM in Azure, but Linux users cannot access its filesystem, 'Expand the correct managed data disk, then extend the guest volume if required' does not satisfy the mechanism that ARM attachment is only the infrastructure step; the guest must recognize and mount the volume correctly. The necessary action is 'Verify guest device discovery, mount point, filesystem and permissions', which should be checked against this Azure resource's actual permissions and state.

The nearby action 'Check disk encryption, region, state and attachment compatibility before migration' is associated with this separate need: an encrypted managed disk is moved for attachment to a different VM. That is because disk availability and key access matter; a new VM cannot always attach arbitrary encrypted disks. Here a disk appears attached to a VM in Azure, but Linux users cannot access its filesystem, so 'Check disk encryption, region, state and attachment compatibility before migration' misses the fact that ARM attachment is only the infrastructure step; the guest must recognize and mount the volume correctly. 'Verify guest device discovery, mount point, filesystem and permissions' is the more defensible action; confirm its behavior in the target service before calling the task complete.

'Extend the guest partition and filesystem into the new capacity' solves another task because cloud disk expansion does not always automatically grow the guest partition or filesystem. For the situation where a disk appears attached to a VM in Azure, but Linux users cannot access its filesystem, this distractor misses the requirement that ARM attachment is only the infrastructure step; the guest must recognize and mount the volume correctly. Unlike 'Extend the guest partition and filesystem into the new capacity', 'Verify guest device discovery, mount point, filesystem and permissions' directly targets that requirement; its success must be confirmed at the authorized resource scope.

The nearby action 'Extend the guest partition and filesystem into the new capacity' is associated with this separate need: an admin enlarged a managed disk in Azure, but the Windows guest still shows the old filesystem size. That is because cloud disk expansion does not always automatically grow the guest partition or filesystem. Here an application writes actively to a data disk that an admin wants to detach, so 'Extend the guest partition and filesystem into the new capacity' misses the fact that detaching an active filesystem risks unflushed data loss and application corruption. 'Quiesce writes, unmount safely and coordinate detach with application owner' is the more defensible action; confirm its behavior in the target service before calling the task complete.

'Verify guest device discovery, mount point, filesystem and permissions' solves another task because ARM attachment is only the infrastructure step; the guest must recognize and mount the volume correctly. For the situation where an application writes actively to a data disk that an admin wants to detach, this distractor misses the requirement that detaching an active filesystem risks unflushed data loss and application corruption. Unlike 'Verify guest device discovery, mount point, filesystem and permissions', 'Quiesce writes, unmount safely and coordinate detach with application owner' directly targets that requirement; its success must be confirmed at the authorized resource scope.

When a database VM's data volume is full while its operating system disk has space, the action 'Expand the correct managed data disk, then extend the guest volume if required' can be relevant: OS and data disks hold different filesystems; adding OS capacity may not fix database storage. For the different problem of an application writes actively to a data disk that an admin wants to detach, 'Expand the correct managed data disk, then extend the guest volume if required' misses the mechanism that detaching an active filesystem risks unflushed data loss and application corruption. Choosing 'Expand the correct managed data disk, then extend the guest volume if required' would address that other need, whereas 'Quiesce writes, unmount safely and coordinate detach with application owner' is intended to solve the stated condition.