ACL Placement and Troubleshooting
Knowing the syntax is half the battle; knowing where to put the list and how to debug it when traffic vanishes is the other half. This lesson turns the placement rules into a decision procedure, adds the VTY-focused access-class twist, and gives you a step-by-step routine for answering “why is everything blocked?” — the question every exam and job interview asks.
- Decide for any scenario whether a standard or extended ACL fits, and where it belongs.
- Trace traffic through an ACL in the correct ordering to find the matching rule.
- Recognise the most common ACL configuration mistakes.
- Restrict router management access with access-class on the VTY lines.
- Verify a live ACL with show access-lists, show ip interface, and the log keyword.
- Follow a methodical troubleshooting sequence for ACL failures.
Standard vs extended: where each belongs
The placement decision is the first question on any ACL scenario. The rules are simple and opposite:
| ACL type | What it can match | Where it goes | Why |
|---|---|---|---|
| Standard | Source address only | Close to the destination | It cannot see the destination, so filtering early would block too much. |
| Extended | Source, destination, protocol, port | Close to the source | It knows the destination, so it can drop unwanted traffic before it crosses the network. |
Two more practical filters help in real networks. Restrict management access by source IP (a standard list on the VTY lines), and filter user data precisely by service (an extended list near the users). The two live in completely different places and both are legitimate per the blueprint.
Testing order: how a packet meets the list
When you (or the exam) need to know what one packet does, run the evaluation in a fixed order:
- Direction. Determine which interface and direction the packet crosses; only the ACL bound there applies.
- Top to bottom. Compare the packet to line 1; if it matches, apply that action and stop.
- No match. If no statement matches anywhere, the implicit deny applies.
Worked example: ACL 101 has permit tcp 10.1.1.0 0.0.0.255 any eq 80, then
deny ip any any. A TCP packet from 10.1.1.5 destined for web port 443 matches line 1’s
protocol and source but the destination port 443 is not equal to 80 — so line 1 does not match and
the packet falls to line 2, which denies it. Many candidates stop at “the source is permitted”
and forget that a non-matching port keeps the packet moving down the list.
The common mistakes
Pick any ACL that “does not work” and one of these is almost always in play:
- Order. A broad
deny anyplaced before the specific permits shadows them. First match wins means reordering is mandatory, not optional. - Missing permit any. Without it, implicit deny silently kills unmatched traffic.
- Wrong direction or interface. The list exists but the traffic never passes it.
Check
show ip interfaceon every interface in the path. - Standard ACL used for port filtering. Illegal syntax, doesn’t filter.
- Extended ACL matching the wrong port.
eqnext to the source matches source ports (random ephemeral values), confusing every permit/deny decision. - Editing a numbered ACL by appending. New numbered lines are always added at the end, so you cannot push a deny ahead of an existing broad permit. Use a named ACL to insert with sequence numbers.
- Forgetting the return path. Blocking inbound TCP to a server does not stop the server’s replies unless the router also filters outbound; a one-direction filter can break conversations asymmetrically.
Putting an ACL on the VTY lines: access-class
VTYs are the router’s remote-management doors. Instead of applying an ACL to an interface, you filter management sessions directly on the lines with access-class, which works like ip access-group but for Telnet/SSH login sources. A standard ACL is typical, because the only fact that matters is the caller’s source address:
! Permit only the management subnet to log in remotely R1(config)# access-list 15 permit 192.168.10.0 0.0.0.255 R1(config)# line vty 0 4 R1(config-line)# access-class 15 in R1(config-line)# transport input ssh R1(config-line)# login local R1(config-line)# exit
ip access-group; VTY lines use
access-class. Mixing the two commands is a revered trap: applying an ACL that was meant
for management with access-group does not restrict logins at all.
Because implicit deny still applies, only sources on 192.168.10.0/24 can reach the VTY lines. If you accidentally lock yourself out, you recover on the console (still open — local, out of the path). That “I just locked myself out” scenario is exactly what the exam wants you to predict from the implicit-deny rule.
Verifying ACLs
Three commands answer nearly every verification question:
| Command | What it shows | When to use it |
|---|---|---|
show access-lists | All ACLs and their per-line hit counters | Is the rule being hit at all? |
show ip interface | Which ACL is bound, in which direction, on each interface | Is the chosen interface/direction right? |
show running-config | The exact configured ACL lines | Catching a typo or bad wildcard |
The workflow: confirm the bind first (does this interface see the traffic?), then the counters (does the packet hit the expected line?), then the order (did a broad permit swallow it?). The counters are the crown jewel: a deny with zero hits means the packet matched something earlier or never arrived here — both are clues, not mysteries.
Logging with the log keyword
Appending log to a statement makes IOS generate a syslog message each time traffic matches it — invaluable for seeing denied attempts live instead of guessing from counters:
R1(config)# access-list 100 deny tcp any host 203.0.113.10 eq 23 log R1(config)# access-list 100 permit ip any any R1# show logging *Oct 12 09:41:07.274: %SEC-6-IPACCESSLOGP: list 100 denied tcp 172.16.8.9(51234) -> 203.0.113.10(23), 1 packet R1# show access-lists Extended IP access list 100 10 deny tcp any host 203.0.113.10 eq telnet log (12 matches) 20 permit ip any any (500 matches)
A syslog entry appears with a timestamp, the source and destination (with ports), and the matched
rule. Two operational cautions: logging every matching packet can flood the CPU — IOS batches and
rate-limits syslog generation, but a heavily denied server can still generate enormous log volume —
and you should only log what you intend to inspect. When a question asks how to see WHICH attacker was
hitting a rule, the answer is the log keyword.
A troubleshooting routine
When “nothing works after I added an ACL”, follow this ladder instead of guessing:
- Confirm the bind.
show ip interface <if>: is the list on this interface, in the intended direction? - Confirm the list.
show access-lists: does the intended line exist, with the right order and syntax? - Confirm the hit. Are counters incrementing on the line you expect? Zero hits means the packet matched above or passes elsewhere.
- Confirm the permit-any. Is there an explicit catch-all, or did implicit deny remove traffic you expected to flow?
- Confirm both directions. Is the reply traffic also permitted, or did you build an asymmetric filter?
- Confirm end-to-end. Run
ping/traceroutefrom the source and test the specific service (the port matters — ICMP can pass while TCP 443 still drops).
show ip interface), a broad permit sits above a specific deny (reorder),
or the final permit any is missing. Walk the ladder and you will spot all three in under a
minute.
The vocabulary you need
Every term below appears in the lesson text above. Select one to read its definition, then use the quiz at the end to check what stuck.
Review
Check what you have learned with these review questions.
Sources & references
- Cisco. CCNA (200-301) Exam Topics. cisco.com.
- IETF. Request for Comments (RFCs). rfc-editor.org.
- IANA. Internet Assigned Numbers Authority registries. iana.org.
Target exams: CCNA 200-301 v1.1 · CyberOps Associate 200-201. Last content verification: 2026-10-11.