Data center: Restore both services without opening the boundary
The incident
Application workers connect to database and connection-pool subnets through separate access and service aggregation layers. Four routed leaves share two spines; clients and services use separate access handoffs while partner and monitoring traffic terminate on other leaves. Both authorized service transports fail after a policy rollout. ICMP still works, and untrusted clients must remain blocked from both services. Investigate database connectivity from R1: 10.168.96.0/24 to 10.168.110.0/24 (TCP/5432), and 10.168.97.0/24 to 10.168.111.0/24 (TCP/6432).
Your objective
Restore both authorized service transports and return reachability, including clients and service hosts in both halves of the primary subnet. Keep R5 blocked from both service transports, keep SSH blocked to both service subnets, and preserve monitoring between R1 and R6. Do not change physical topology or addressing. These are network forwarding checks, not application-level transactions.
The network
Link addressing (10)
- R1 R2
10.168.112.0/30 - R2 R3
10.168.113.0/30 - R2 R4
10.168.114.0/30 - R5 R3
10.168.115.0/30 - R5 R4
10.168.116.0/30 - R6 R3
10.168.117.0/30 - R6 R4
10.168.118.0/30 - R7 R3
10.168.119.0/30 - R7 R4
10.168.120.0/30 - R7 R8
10.168.121.0/30
What you’ll practice
Interview exercises are inspired by common industry competencies. They are not proprietary interview questions.