Google IT Support cheat sheet

Six-step troubleshooting model: identify the problem, establish a theory, test the theory, plan and execute the fix, verify full functionality, and document findings.

Open-ended questions gather symptom detail; closed-ended questions confirm specific facts after the problem space is narrowed.

Change one variable at a time so the effect of each modification can be attributed to a single cause.

Escalate when the issue involves safety, data-loss risk, missing permissions, or knowledge beyond the current support tier.

A useful ticket record contains symptoms, affected user and device, tests performed, root cause, resolution, and follow-up needed.

Paraphrasing and summarizing the user's report confirms understanding and exposes gaps before technical work begins.

Set and restate expectations for next steps and response time, then confirm the user agrees before closing.

OSI model layers: physical, data link, network, transport, session, presentation, application.

Private IPv4 ranges: 10.0.0.0/8, 172.16.0.0/12, and 192.168.0.0/16.

A 169.254.x.x address is APIPA and typically means DHCP failed.

DNS resolves names to IP addresses; DHCP assigns IP address, subnet mask, gateway, and DNS servers.

The default gateway is used for destinations outside the local subnet.

TCP is reliable and connection-oriented; UDP is connectionless and does not guarantee delivery.

Common ports: 22 SSH, 25 SMTP, 53 DNS, 80 HTTP, and 443 HTTPS.

A /24 IPv4 subnet has 256 addresses, 254 usable hosts, one network address, and one broadcast address.

ping tests ICMP reachability; traceroute shows the path and latency per hop.

NTFS supports Windows permissions, encryption, compression, journaling, and large files; FAT32 has a 4 GB maximum file size.

Linux permissions use r=4, w=2, x=1; 755 means rwxr-xr-x.

Windows Safe Mode loads minimal drivers and services to isolate third-party software causes.

Boot order is firmware POST, bootloader, kernel, then user-space services.

Task Manager on Windows, Activity Monitor on macOS, and top or ps on Linux show running processes and resource use.

A standard user account cannot make system-wide changes without elevation; an administrator account can.

Virtual memory or swap extends RAM using disk and is much slower than physical memory.

File paths use backslashes on Windows and forward slashes on Linux and macOS.

Least privilege grants only the access required for a role or task.

The 3-2-1 backup rule: three copies, two media types, one off-site.

RAID 0 stripes for performance with no redundancy; RAID 1 mirrors; RAID 5 uses distributed parity and tolerates one drive failure.

Change management requires request, review, test, approval, deployment, and documentation.

Patch management identifies, tests, deploys, and verifies updates.

Account lifecycle includes onboarding, role changes, and timely offboarding with access removal.

Monitoring covers CPU, memory, disk capacity, network, logs, and service availability.

SSH provides encrypted remote command-line access; RDP provides remote Windows desktop access.

The CIA triad is confidentiality, integrity, and availability.

MFA combines factors from knowledge, possession, and inherence categories.

Phishing indicators include urgency, mismatched sender domains, unexpected attachments, and credential requests.

Encryption in transit uses TLS; encryption at rest protects stored data.

A firewall permits or blocks traffic based on rules such as source, destination, port, and protocol.

Password hashing is one-way; salting adds unique data per password to defeat precomputed tables.

Malware categories include virus, worm, Trojan, ransomware, spyware, and adware.

Social engineering targets people through pretexting, baiting, tailgating, and impersonation.

AI outputs are probabilistic and can be confidently wrong; verify facts, citations, and calculations against trusted sources.

A specific prompt states goal, audience, format, constraints, and examples.

Do not enter confidential, personal, or regulated data into unapproved AI tools.

Generative AI is effective for outlining, summarizing, rewriting, brainstorming, and translating drafts.

Hallucination is fabricated or incorrect content presented as fact.

Bias in AI can come from training data, labeling, or design and requires testing and human review.

AI-generated code, text, and images may have licensing or copyright constraints.

Keep a human decision-maker accountable for final outputs used in support, security, or administration.