Quiz — Networking Production¶
Quiz Overview¶
Assess Module 7 Production Network Operations: segmentation, DNS ops, load-balancer health checks, firewall change control, and incident response.
| Attribute | Value |
|---|---|
| Topic | Production network operations |
| Questions | 25 |
| Passing score | 70% (18 correct) |
| Estimated time | 30–40 minutes |
How to use this quiz
Attempt without peeking. Use Reveal answer only after you choose.
Learning Objectives¶
- Reason about trust zones, blast radius, and tier allow matrices
- Plan DNS TTL cutovers and split-horizon behaviour
- Operate L4/L7 health checks, drain, and sticky-session trade-offs
- Apply firewall change control with rollback and console safety
- Execute first-hour incident evidence collection with curl, dig, and ss
Section 1 — Fundamentals¶
Question 1¶
Which trust zone typically terminates public HTTPS and sits closest to the internet?
Difficulty: Beginner
Options:
- A. Edge tier (load balancers, WAF, CDN)
- B. Data tier (databases)
- C. Management tier (bastion hosts)
- D. Application tier (workers only)
Reveal answer
Correct answer: A
Edge zones accept untrusted inbound traffic and forward to app tiers under policy.
Related concepts
- Trust zones
- Segmentation
Question 2¶
What does blast radius measure in network segmentation?
Difficulty: Beginner
Options:
- A. Maximum cable length in a rack
- B. DNS TTL in seconds
- C. How far compromise or misconfiguration can spread across tiers
- D. HAProxy
maxconnsetting
Reveal answer
Correct answer: C
Segmentation shrinks blast radius by limiting lateral movement.
Related concepts
- Blast radius
- Trust boundaries
Question 3¶
Before lowering DNS TTL for a migration, you must typically wait because?
Difficulty: Intermediate
Options:
- A. TTL changes propagate instantly worldwide
- B. Authoritative servers reject low TTL always
- C. curl ignores TTL
- D. Resolvers must expire the previous TTL before caching the new value
Reveal answer
Correct answer: D
Old cached answers persist until their original TTL expires — plan 24–48 hours ahead.
Related concepts
- TTL
- DNS caching
Question 4¶
Split-horizon DNS means?
Difficulty: Intermediate
Options:
- A. One hostname always resolves to the same global IP
- B. Different answers for the same name depending on client location or resolver view
- C. DNS records stored only in
/etc/hosts - D. IPv6-only resolution
Reveal answer
Correct answer: B
Internal and external views may differ — test both VPN and public paths.
Related concepts
- Split-horizon
- DNS operations
Question 5¶
An L4 (TCP) health check validates?
Difficulty: Beginner
Options:
- A. TCP connection establishment to a port
- B. HTTP status code and response body
- C. TLS certificate expiry only
- D. SQL query success
Reveal answer
Correct answer: A
TCP connect confirms the port accepts — not that the application logic is healthy.
Related concepts
- L4 health checks
- Load balancing
Question 6¶
An L7 HTTP health check can detect which failure mode that TCP alone often misses?
Difficulty: Intermediate
Options:
- A. Port closed by firewall only
- B. Wrong DNS A record
- C. Process listening but returning HTTP 503 on
/healthz - D. Cable unplugged
Reveal answer
Correct answer: C
HTTP probes validate application-layer readiness, not just socket accept.
Related concepts
- L7 health checks
- httpchk
Question 7¶
Drain on a load-balancer backend means?
Difficulty: Intermediate
Options:
- A. Immediately kill all connections to that node
- B. Delete the DNS record
- C. Format the disk
- D. Stop sending new connections; allow existing sessions to complete
Reveal answer
Correct answer: D
Drain enables zero-downtime maintenance windows.
Related concepts
- Drain
- Graceful shutdown
Question 8¶
Sticky sessions (session affinity) primarily trade what operational cost?
Difficulty: Advanced
Options:
- A. Faster TLS handshakes always
- B. Uneven load distribution and lost sessions when a node fails without external state
- C. Lower DNS query volume
- D. Automatic firewall rollback
Reveal answer
Correct answer: B
Prefer external session stores over stickiness for new systems.
Related concepts
- Sticky sessions
- Load distribution
Section 2 — Practical Knowledge¶
Question 9¶
In HAProxy with inter 2s fall 3, approximately how long until a failing backend is marked DOWN?
Difficulty: Intermediate
Options:
- A. About 6 seconds (three failed checks at 2s interval)
- B. Instantly on first failed check
- C. Exactly 24 hours
- D. Never — TCP checks only
Reveal answer
Correct answer: A
fall 3 requires three consecutive failures at inter 2s spacing.
Related concepts
- HAProxy health checks
- inter/fall/rise
Question 10¶
After killing one backend in a two-node HAProxy pool with working health checks, user traffic should?
Difficulty: Beginner
Options:
- A. Fail entirely until manual intervention
- B. Round-robin to the dead node anyway
- C. Route only to surviving healthy backends
- D. Switch to UDP
Reveal answer
Correct answer: C
Failed nodes are removed from rotation; survivors absorb load.
Related concepts
- Failover
- HAProxy backend state
Question 11¶
Why should you confirm console or serial access before applying restrictive host firewall rules?
Difficulty: Advanced
Options:
- A. Console speeds up DNS TTL
- B. ufw requires serial port licensing
- C. HAProxy only runs on serial consoles
- D. A bad deny rule can lock out SSH — console is the rollback path
Reveal answer
Correct answer: D
Out-of-band access prevents permanent lockout from ACL mistakes.
Related concepts
- Console access
- Firewall change control
Question 12¶
Before applying a firewall change, capturing sudo nft list ruleset or ufw status numbered supports?
Difficulty: Intermediate
Options:
- A. Faster round-robin
- B. Documented rollback to the pre-change ruleset
- C. Automatic DNS failover
- D. Sticky session removal
Reveal answer
Correct answer: B
Export current rules before change — rollback values must exist before you need them.
Related concepts
- ACL rollback
- Change control
Question 13¶
Lowering DNS TTL to 300 seconds before a cutover primarily helps?
Difficulty: Intermediate
Options:
- A. Shorten cache lifetime so rollback or new IPs propagate faster after the change
- B. Increase steady-state resolver load forever
- C. Disable split-horizon
- D. Replace authoritative nameservers
Reveal answer
Correct answer: A
Lower TTL during migration windows reduces stale-cache pain.
Related concepts
- TTL planning
- DNS cutover
Question 14¶
A tier allow matrix documents?
Difficulty: Beginner
Options:
- A. Only usernames and passwords
- B. HAProxy algorithm names only
- C. Source zone, destination zone, port, protocol, and business justification
- D. Certificate SAN lists only
Reveal answer
Correct answer: C
Matrices make segmentation auditable and reviewable.
Related concepts
- Tier allow matrix
- Segmentation
Section 3 — Scenarios & Operations¶
Question 15¶
Friday 17:00: a new deny rule blocks TCP 443 on the edge LB. Synthetic checks fail globally. First stabilisation step?
Difficulty: Advanced
Options:
- A. Reboot all application servers
- B. Raise DNS TTL to 86400
- C. Enable sticky sessions
- D. Roll back or remove the recent ACL change after confirming scope
Reveal answer
Correct answer: D
Assume recent change until proven otherwise — rollback ACL with documented pre-change export.
Related concepts
- Incident response
- ACL rollback
Question 16¶
During a deploy, ops sets backend weight to 0 and waits for active sessions to finish before restart. This is?
Difficulty: Intermediate
Options:
- A. Hard failover
- B. Graceful drain
- C. Split-horizon cutover
- D. TTL flush
Reveal answer
Correct answer: B
Drain prevents dropping in-flight user requests during maintenance.
Related concepts
- Drain
- Deploy windows
Question 17¶
VPN users resolve app.internal.example.com to 10.0.1.50; public resolvers return 203.0.113.10 for app.example.com. This illustrates?
Difficulty: Intermediate
Options:
- A. Split-horizon / split DNS views
- B. Flat network design
- C. HAProxy round-robin
- D. TCP health check only
Reveal answer
Correct answer: A
Internal and external names or views differ by design — test both.
Related concepts
- Split-horizon
- Production DNS
Question 18¶
A canary ACL change should be applied first on?
Difficulty: Advanced
Options:
- A. Every production host simultaneously
- B. The database primary only
- C. A single non-critical host or port with observation before fleet rollout
- D. Public internet root servers
Reveal answer
Correct answer: C
Smallest blast radius: canary one target, validate, then promote via IaC.
Related concepts
- Canary ACL
- Blast radius
Question 19¶
App tier should accept traffic from which source in a standard three-tier web design?
Difficulty: Beginner
Options:
- A. Any IP on the internet directly to port 8080
- B. Database tier only
- C. Management bastion on all ports
- D. Edge/load-balancer tier CIDR or security group only
Reveal answer
Correct answer: D
App tier trusts edge, not the raw internet — segmentation enforces the hop.
Related concepts
- Trust zones
- Tier matrix
Section 4 — Troubleshooting¶
Question 20¶
First-hour network incident evidence collection should start with which command set?
Difficulty: Beginner
Options:
- A.
rm -rf /then reboot - B. Structured
curl,dig, andssoutput saved with UTC timestamps - C. Random log grep with no timestamps
- D. Disable all monitoring
Reveal answer
Correct answer: B
curl proves HTTP path; dig proves name resolution; ss proves listeners and connections.
Related concepts
- Incident response
- Evidence toolkit
Question 21¶
curl to 127.0.0.1:18080 returns 200, but curl to edge.rebash.lab:18080 fails. getent shows 127.0.0.99. Likely layer?
Difficulty: Intermediate
Options:
- A. Name resolution / hosts override fault
- B. Application crash
- C. TLS certificate expiry
- D. LVM full
Reveal answer
Correct answer: A
IP works, FQDN fails — suspect DNS or /etc/hosts before the app.
Related concepts
- DNS triage
- getent vs curl
Question 22¶
ss shows HAProxy listening on 18080, but curl to the edge times out. Firewall was changed recently. Likely cause?
Difficulty: Intermediate
Options:
- A. HAProxy not installed
- B. Wrong HTTP health-check path only
- C. ACL or security group blocking the path despite process listening
- D. DNS TTL too low
Reveal answer
Correct answer: C
Listener present + connection failure → path/ACL or routing, not "process down".
Related concepts
- ss vs curl
- Firewall triage
Question 23¶
dig @8.8.8.8 returns empty for an internal hostname, but getent hosts resolves it. Best explanation?
Difficulty: Advanced
Options:
- A. dig is broken
- B. HAProxy removed the backend
- C. TTL is infinite on the internet
- D. Public recursive DNS has no record; libc resolver uses
/etc/hostsor internal stub
Reveal answer
Correct answer: D
dig queries configured resolvers; getent honours local hosts — always check both.
Related concepts
- dig vs getent
- Split-horizon
Section 5 — Architecture¶
Question 24¶
In production, why bind application backends to 127.0.0.1 behind a load balancer?
Difficulty: Intermediate
Options:
- A. To bypass health checks
- B. To keep the app off the public exposure path; only the LB front door is reachable
- C. Because Python requires localhost
- D. To disable round-robin
Reveal answer
Correct answer: B
Localhost bind limits blast radius — clients must traverse the LB tier.
Related concepts
- Exposure control
- Edge architecture
Question 25¶
Module 7 production network operations integrates which concerns holistically?
Difficulty: Advanced
Options:
- A. Segmentation, DNS change control, LB health checks, ACL rollbacks, and incident timelines
- B. Only packet capture with tcpdump
- C. Only LVM and fstab
- D. Container image scanning only
Reveal answer
Correct answer: A
Module 7 ties trust boundaries through to first-hour incident evidence and rollback discipline.
Related concepts
- Module 7 capstone
- Production operations
Score Summary¶
| Score | Band | Meaning |
|---|---|---|
| 23–25 | Excellent | Ready for edge failover lab and networking interviews |
| 18–22 | Good | Pass — revise weak Module 7 tutorials |
| 12–17 | Needs improvement | Revisit tutorials 21–25 and foundation DNS/firewall labs |
| ≤11 | Restart Module 7 | Work tutorials 21–25 in order |
Total questions: 25 · Passing score: 18 (70%)
Recommended Study Areas¶
- Networking Module 7
- Lab: Networking Edge Failover
- Lab: DNS and Firewall Site-Down Triage
- Cheat sheet: Networking
- Interview: Networking
Interview Connection¶
Expect trust-zone design, TTL cutover planning, L4 vs L7 health checks, drain vs failover, ACL rollback stories, and first-hour curl/dig/ss triage. Use your Module 7 lab narrative.