At 9:00 am each morning, all of the virtual desktops in a VDI implementation become extremely slow and/or unresponsive. The outage lasts for around 10 minutes, after which everything runs properly…

Question

Options

• AAdd guests with more memory to increase capacity of the infrastructure.
• BA backup is running on the thin clients at 9am every morning.
• CInstall more memory in the thin clients to handle the increased load while booting.
• DBooting all the lab desktops at the same time is creating excessive I/O.
• EInstall 10-Gb uplinks between the hosts and the lab to increase network capacity.
• FInstall faster SSD drives in the storage system used in the infrastructure.
• GThe lab desktops are saturating the network while booting.
• HThe lab desktops are using more memory than is available to the host systems.

• A
  9% (4)
• B
  2% (1)
• C
  4% (2)
• D
  63% (29)
• E
  20% (9)
• G
  2% (1)

Explanation

This is a classic 'boot storm' scenario. When all thin clients boot simultaneously at 9:00 AM, each virtual desktop instance must load its OS, applications, and profile from shared storage all at once. This creates a massive spike in storage I/O that overwhelms the storage subsystem, causing all VMs to become slow or unresponsive until the burst subsides (~10 minutes). D correctly identifies the root cause: simultaneous booting creates excessive I/O on the shared storage. F correctly identifies the best solution: installing faster SSD drives in the storage infrastructure dramatically increases IOPS capacity, allowing the system to handle the burst I/O load without degrading performance. Adding memory (A, C) or network upgrades (E) do not address the storage I/O bottleneck that is the actual constraint.

No community discussion yet for this question.