Lab 5 — Troubleshooting Drill
Goal: rehearse the scenarios under time pressure, out loud, against your own lab. Time: ~2 hours.
Reading about troubleshooting does not help under interview conditions. Only doing it, out loud, with a clock, helps. Set a timer for each scenario.
How to run each scenario
- Set a 5 minute timer.
- Break the environment using only commands you know.
- Narrate your thinking out loud as if someone were watching — this is the part interviewers actually score.
- Fix it.
- Note the commands you reached for and the ones you forgot.
- Restore from snapshot.
Scenario 1 — Locked account (5 min)
Break:
net user ravi /active:no
Narrate this:
"First, scope — is it just this user or everyone? One user points at the account. I'd ask for the exact error text. Then
net user ravito see whether it's locked, disabled or expired. If locked,net user ravi /active:yes, then reset the password withnet user ravi *. If the error mentions trust or domain I'd checknltest /dsgetdc:corp.local. I'd also resync time withw32tm /resync, because clock skew over five minutes breaks Kerberos. And I'd ask them to test on the LAN — if they can log in offline, their credentials are cached and the account is fine."
Reset: net user ravi /active:yes
Scenario 2 — DHCP exhaustion (5 min)
Break:
Set-DhcpServerv4Scope -ScopeId 192.168.10.0 -LeaseDuration 00:02:00
Force repeated renews until the client fails.
Narrate:
"169.254 address means the DHCP request failed, so this isn't a connectivity problem — the client is on the network, it just got no lease. I'd check the scope on the server: lease statistics, whether the scope is active, and what interface it's bound to. A very short lease makes exhaustion obvious, but in production I'd also suspect the lease duration is too long for the number of users."
Scenario 3 — DNS record deleted (4 min)
Break:
Remove-DnsServerResourceRecord -ZoneName "corp.local" -Name "srv01" -Type A -Force
Narrate:
"Ping by IP works but ping by name fails, which tells me immediately this is name resolution rather than connectivity — different layers. I'd
nslookupto confirm, checkipconfig /allto make sure the client is using the right DNS server, flush the client cache withipconfig /flushdns, then query the authoritative server directly. If the authoritative server also returns nothing, it's a bad record rather than a bad client."
Scenario 4 — GPO not applying (6 min)
Break:
Move-ADObject -Identity "CN=SRV01,CN=Computers,DC=corp,DC=local" `
-TargetPath "OU=HR,OU=CORP,DC=corp,DC=local"
Narrate:
"I'd check where the object actually is first —
Get-ADComputer. Thengpresult /r, which gives me the resultant set plus the reason for each policy. Here it will say the policy wasn't applied because the object is outside the OU's scope. I'd move it back, force withgpupdate /force, and confirm with a second report."
Scenario 5 — Print server down (4 min)
Break:
Stop-Service Spooler -ComputerName "\\print01"
Narrate:
"I'd clarify scope first — does anything print, anywhere, or is it one machine? Then check the physical state, look at the queue in PrintManagement for a specific error, test
Test-NetConnectionto the printer on port 9100, and check whether the spooler service is running. If the queue is stuck deleting, that's the spooler, and restarting it is the fix."
Scenario 6 — Whole floor offline (6 min)
Break: disconnect the pfSense LAN interface, or disable a switch port.
Narrate:
"Scope says infrastructure, so I wouldn't touch a workstation. I'd check the switch uplink state and look for errors, check whether it's one VLAN or the whole floor, and check the firewall. I would not reboot the uplink — it affects everyone and destroys the state that tells me what happened. If it came back after about thirty seconds I'd be looking at spanning tree."
Self-scoring
Rate yourself honestly out of 5 for each:
| Dimension | 1 | 3 | 5 |
|---|---|---|---|
| Scope question asked first | |||
| Correct layer identified | |||
| Right command, no notes | |||
| Explained the reasoning aloud | |||
| Verified with the user |
Repeat the whole set every week until you can do each in under five minutes without notes. That is the level you need in an interview.
Verification
You're done
Go back to Technical Questions and rehearse the written answers. Then read STAR Stories & Resume and rewrite your resume for this job.
You built the lab. That was the hard part.