Policy-Based Routing
Policy routing by source, marks or alternate tables, multi-WAN design, failover and asymmetry troubleshooting.
What this topic solves
Policy routing by source, marks or alternate tables, multi-WAN design, failover and asymmetry troubleshooting.
Follow state and packet path
Start with next-hop reachability, RIB/FIB and route source. Then validate protocol state, policy, metrics, attributes and convergence.
Input → classification/state → decision → table → next-hop/output → countersDesign before configuration
- Define the failure domain and expected convergence behavior.
- Document addressing, identifiers, timers, MTU and policy boundaries.
- Prefer explicit, observable state over assumptions.
- Plan rollback and out-of-band access before risky changes.
Commands and verification
Command syntax varies by platform and software release. The examples below come from the existing portal content and must be validated against the target platform documentation before production use.
ip access-list extended CLIENTES-A
permit ip 10.10.10.0 0.0.0.255 any
route-map WAN2 permit 10
match ip address CLIENTES-A
set ip next-hop 192.0.2.2
interface GigabitEthernet0/1
ip policy route-map WAN2
show route-map
show ip policy
show access-lists
/routing/table/print
/routing/rule/print detail
/ip/firewall/mangle/print detail stats
/ip/route/print detail
# Fluxo conceitual:
# classificar → mark-routing → tabela alternativa
# ou usar /routing/rule conforme desenho.Evidence-driven workflow
- Define source, destination, time window and exact symptom.
- Establish the expected healthy state and compare it with the observed state.
- Inspect counters, tables, logs and packet captures before resets.
- Test one hypothesis at a time and prefer reversible changes.
- Repeat the original test after correction and check side effects.
Standards and RFCs applicable to the technology
Use the current standard/RFC and the documentation for the exact software release being operated. Vendor syntax is an implementation detail; protocol behavior should be understood independently.
Prove the behavior in a controlled topology
Build the smallest topology that reproduces the mechanism, capture control and data traffic, introduce one deliberate failure and diagnose it before applying the fix.