Trace a Computer's Boot From the Power Button to a Running OS

Quick Overview

A security-focused SRE internship question asking the candidate to trace a computer's startup from pressing the power button to a running operating system. It tests firmware, boot selection, bootloader, kernel and init knowledge, plus the boot chain of trust and diagnosing which stage failed.

Trace a Computer's Boot From the Power Button to a Running OS

Company: ByteDance

Role: Site Reliability Engineer

Category: Software Engineering Fundamentals

Difficulty: easy

Interview Round: Technical Screen

You are interviewing for a security-focused site reliability engineering internship. The interviewer asks: what happens from the moment you press the power button on a computer until the operating system is up and running? Walk through every stage in order, and expect the interviewer to stop you at any stage and ask for more detail. ```hint Follow who is in control At each stage, ask which piece of code is running, where it was loaded from, and what it hands control to next. ``` ```hint The security angle For a security team, consider how each stage could know that the next stage it loads has not been tampered with. ``` ### Clarifying Questions - Should the answer assume a typical x86-64 machine with UEFI firmware running Linux, or also cover legacy BIOS, ARM devices or Windows? - Is the machine a laptop or desktop booting from a local disk, or a server that may boot over the network? - Should the answer stop when the kernel starts the first user-space process, or continue until services are running and a user can log in? - Is the root disk encrypted? ### What a Strong Answer Covers - The hardware start: power stabilizing, the CPU leaving reset and fetching its first instruction from firmware - Firmware duties: initializing memory and devices, the power-on self-test, and choosing a boot device (UEFI boot entries versus a legacy master boot record) - The bootloader's job: loading the kernel and the initial RAM disk and passing the kernel command line - Kernel initialization through mounting the real root filesystem and starting the first user-space process - The init system bringing up services and a login prompt - The chain of trust: Secure Boot signature checks, measured boot with a TPM, and unlocking an encrypted disk - Where boot can fail at each stage, and how a symptom tells you which stage failed ### Follow-up Questions - How does Secure Boot stop a modified bootloader or kernel from running, and what does measured boot add that Secure Boot does not? - The machine stops with a kernel panic saying it cannot mount the root filesystem. Which stages worked, and what would you check first? - Why does a Linux boot usually need an initial RAM disk at all? - How does a server in a data center boot when there is no operating system on its local disk?

Overview: A security-focused SRE internship question asking the candidate to trace a computer's startup from pressing the power button to a running operating system. It tests firmware, boot selection, bootloader, kernel and init knowledge, plus the boot chain of trust and diagnosing which stage failed.

ByteDance logo
ByteDance
Apr 9, 2026
easySite Reliability EngineerTechnical ScreenSoftware Engineering Fundamentals
0
0

You are interviewing for a security-focused site reliability engineering internship. The interviewer asks: what happens from the moment you press the power button on a computer until the operating system is up and running? Walk through every stage in order, and expect the interviewer to stop you at any stage and ask for more detail.

Clarifying Questions Guidance

  • Should the answer assume a typical x86-64 machine with UEFI firmware running Linux, or also cover legacy BIOS, ARM devices or Windows?
  • Is the machine a laptop or desktop booting from a local disk, or a server that may boot over the network?
  • Should the answer stop when the kernel starts the first user-space process, or continue until services are running and a user can log in?
  • Is the root disk encrypted?

What a Strong Answer Covers Guidance

  • The hardware start: power stabilizing, the CPU leaving reset and fetching its first instruction from firmware
  • Firmware duties: initializing memory and devices, the power-on self-test, and choosing a boot device (UEFI boot entries versus a legacy master boot record)
  • The bootloader's job: loading the kernel and the initial RAM disk and passing the kernel command line
  • Kernel initialization through mounting the real root filesystem and starting the first user-space process
  • The init system bringing up services and a login prompt
  • The chain of trust: Secure Boot signature checks, measured boot with a TPM, and unlocking an encrypted disk
  • Where boot can fail at each stage, and how a symptom tells you which stage failed

Follow-up Questions Guidance

  • How does Secure Boot stop a modified bootloader or kernel from running, and what does measured boot add that Secure Boot does not?
  • The machine stops with a kernel panic saying it cannot mount the root filesystem. Which stages worked, and what would you check first?
  • Why does a Linux boot usually need an initial RAM disk at all?
  • How does a server in a data center boot when there is no operating system on its local disk?