Back to EIGRP GuideENARSI convergence repair

EIGRP SIA and Query Boundaries

Understand why a route becomes Active, how SIA-Query and SIA-Reply keep a slow calculation alive, and how to find the neighbor that did not complete the query process.

Active RouteQueryReplySIAQuery Boundary

SIA at a Glance

A route goes Active when its successor fails and no feasible successor is ready. The router sends Queries and must receive every required Reply before DUAL can finish.

Healthy steady statePassive

Recalculation stateActive

Progress checkSIA-Query/Reply

Design goalBound the query domain

Important: Passive means stable, not disabled. Active for a short time can be normal during convergence.

1. Follow the Active Process

Successor fails

The best path is lost and no feasible successor exists.

Queries spread

Neighbors search their own topology information and may query farther.

Replies return

The originator waits for all required Replies before choosing a result.

2. Understand SIA-Query and SIA-Reply

If a normal Reply is delayed, the querying router uses SIA-Query to ask whether the neighbor is still working on the route. SIA-Reply confirms progress and helps keep the relationship from being reset unnecessarily.

Packet	Purpose	What it proves
Query	Ask for an alternate path	The prefix is Active
Reply	Return an answer	The neighbor completed its work
SIA-Query	Check progress	The original Reply is delayed
SIA-Reply	Confirm work continues	The peer remains responsive

3. Recognize Common Root Causes

A link or neighbor fails inside a large query domain.
A slow, overloaded, or memory-constrained router cannot process Queries promptly.
Packet loss delays reliable EIGRP packets or acknowledgments.
A remote router is itself waiting on another downstream Reply.
Missing summaries or stub boundaries allow Queries to travel too far.

SIA is usually a symptom of the path and query domain, not simply a timer problem.

4. Find the Waiting Neighbor

show ip eigrp topology active
show ip eigrp topology 10.50.0.0/16
show ip eigrp neighbors
show ip eigrp traffic
show interfaces counters errors

In active topology output, look for the peers from which a Reply is still outstanding. Then inspect that direction rather than restarting every router.

5. Build Effective Query Boundaries

Boundary	How it helps	Design check
Feasible successor	Avoids Queries when the successor fails	Does the alternate satisfy feasibility?
EIGRP stub	Stops needless transit Queries at branches	Is the router truly non-transit?
Manual summary	Hides component failures beyond the summary	Can the summarizer reach all advertised space?
Focused topology	Reduces the number of routers involved	Are failure domains intentional?

6. Use a Safe Repair Workflow

Identify the exact Active prefix and originator.
List every neighbor with an outstanding Reply.
Check loss, queues, CPU, and topology in that direction.
Restore the failed link or responsive control plane.
Verify the route returns to Passive and the adjacency remains stable.
After recovery, improve stub, summary, or backup-path design.

Avoid reflexive resets: Clearing EIGRP can remove useful evidence and disrupt healthy routes. Capture state first.

7. Interpret the Evidence

Evidence	Interpretation	Next step
One outstanding peer	Problem is likely in that direction	Inspect peer and downstream chain
High queue count	Reliable packets are waiting	Check loss, congestion, and CPU
Many prefixes Active	Broad failure or control-plane stress	Find common link or neighbor
Frequent adjacency resets	SIA or transport instability	Correlate logs and interface counters

8. Guided Practice

Observe healthy convergence

Record Passive topology.
Fail a path with a feasible successor.
Confirm immediate recovery.

Observe a query

Remove the feasible successor.
Fail the active path.
Follow Query and Reply state.

Contain the domain

Configure a branch as stub.
Repeat the failure.
Compare involved neighbors.

9. Frequently Asked Questions

Is an Active EIGRP route always a problem?

No. A brief Active state is normal when DUAL must ask neighbors for another path.

What makes a route Stuck in Active?

The router does not receive every required response within the allowed Active process.

Should I increase timers to fix SIA?

Usually not first. Find loss, slow routers, long query chains, and missing boundaries before masking the symptom.

What is the best prevention?

Use feasible successors, EIGRP stub at non-transit branches, safe summaries, and a reliable control plane.