InterfacesCCNA 200-301

Output Drops on Interface — Queue Full Under Load

Presenting Symptom

The network engineer observes output drops incrementing on a GigabitEthernet interface under load, causing packet loss and degraded application performance.

Network Context

This issue occurs in a small branch office with a Cisco Catalyst 2960X switch running IOS 15.2, connecting a single server to the corporate WAN via a 1 Gbps link. The interface is configured as a trunk carrying multiple VLANs, and the server generates bursts of traffic exceeding the interface's queue capacity.

Diagnostic Steps

1

Check interface statistics for output drops

show interfaces GigabitEthernet0/1
GigabitEthernet0/1 is up, line protocol is up
  Hardware is Gigabit Ethernet, address is aaaa.bbbb.cccc (bia aaaa.bbbb.cccc)
  MTU 1500 bytes, BW 1000000 Kbit/sec, DLY 10 usec,
     reliability 255/255, txload 1/1, rxload 1/1
  Encapsulation ARPA, loopback not set
  Keepalive set (10 sec)
  Full Duplex, 1000Mbps, media type is 10/100/1000BaseTX
  input flow-control is off, output flow-control is unsupported
  ARP type: ARPA, ARP Timeout 04:00:00
  Last input 00:00:00, output 00:00:00, output hang never
  Last clearing of "show interface" counters never
  Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 12345
  Queueing strategy: fifo
  Output queue: 0/40 (size/max)
  5 minute input rate 500000000 bits/sec, 50000 packets/sec
  5 minute output rate 500000000 bits/sec, 50000 packets/sec
     123456789 packets input, 987654321 bytes, 0 no buffer
     Received 0 broadcasts (0 IP multicasts)
     0 runts, 0 giants, 0 throttles
     0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored
     0 watchdog, 0 multicast, 0 pause input
     123456789 packets output, 987654321 bytes, 0 underruns
     12345 output errors, 12345 collisions, 0 interface resets
     0 unknown protocol drops
     0 babbles, 0 late collision, 0 deferred
     0 lost carrier, 0 no carrier, 0 pause output
     0 output buffer failures, 0 output buffers swapped out

Look for 'Total output drops' and 'Output queue: size/max'. A non-zero output drops count and a queue size near max (e.g., 40/40) indicate the interface is dropping packets because the transmit ring is full. Normal output drops should be zero or very low.

2

Examine the interface queue configuration and current queue depth

show queueing interface GigabitEthernet0/1
Interface GigabitEthernet0/1 queueing strategy:  fifo
  Output queue size (queue depth) = 40
  Output queue max size = 40
  Output queue drops = 12345

This confirms the queue is using FIFO (first-in, first-out) with a maximum depth of 40 packets. The queue depth is at max (40) and drops are high, indicating the queue is consistently full. This suggests the interface is oversubscribed and needs either a larger queue or a more sophisticated queuing strategy.

3

Check if the interface is experiencing congestion by examining traffic rates

show interfaces GigabitEthernet0/1 accounting
Protocol    Pkts In   Chars In   Pkts Out  Chars Out
IP          50000000  4000000000 50000000  4000000000
Other       0         0          0         0

This shows the total packets and bytes for each protocol. If output rates are near line rate (1 Gbps) consistently, the interface is congested. In this scenario, the output rate is 500 Mbps, which is below line rate but bursts may exceed queue capacity.

4

Verify if any QoS policy is applied that could be shaping or policing traffic

show policy-map interface GigabitEthernet0/1
GigabitEthernet0/1

  Service-policy output: (none)

  Service-policy input: (none)

No QoS policy is applied. Without QoS, the interface uses default FIFO queuing, which offers no traffic prioritization or congestion management. This confirms the drops are due to simple tail drop when the FIFO queue fills.

Root Cause

The interface is configured with default FIFO queuing and a small transmit ring (40 packets). Under load, bursts of traffic fill the queue, causing tail drops. The switch has no QoS policy to prioritize critical traffic or manage congestion, leading to packet loss.

Resolution

Apply a QoS policy to prioritize traffic and manage congestion. For example, create a class-map for critical traffic (e.g., VoIP) and apply a policy-map with LLQ (Low Latency Queuing) and fair-queue for other traffic. 1. Create class-map for critical traffic: class-map match-any CRITICAL match ip dscp ef 2. Create policy-map: policy-map QOS-OUT class CRITICAL priority percent 10 class class-default fair-queue 3. Apply to interface: interface GigabitEthernet0/1 service-policy output QOS-OUT Alternatively, if QoS is not desired, increase the transmit ring size: interface GigabitEthernet0/1 tx-ring-limit 2000 (Note: This may not be available on all platforms; check documentation.)

Verification

After applying the QoS policy, verify output drops stop incrementing: show interfaces GigabitEthernet0/1 | include drops Output queue: 0/1000 (size/max); Total output drops: 12345 (The drop count should remain the same or increase very slowly.) show policy-map interface GigabitEthernet0/1 GigabitEthernet0/1 Service-policy output: QOS-OUT Class-map: CRITICAL (match-any) 123456 packets, 987654 bytes 5 minute offered rate 10000000 bps, drop rate 0 bps Match: ip dscp ef (46) Priority: Queue depth 0 packets (no drops) Class-map: class-default (match-any) 1234567 packets, 987654321 bytes 5 minute offered rate 490000000 bps, drop rate 0 bps Match: any Fair-queue: per-flow queue limit 16 Queue depth 0 packets (no drops)

Prevention

1. Implement QoS policies on all WAN and trunk interfaces to prioritize critical traffic and manage congestion. 2. Monitor interface utilization and queue depths proactively using SNMP or NetFlow to detect congestion before drops occur. 3. Design network capacity to handle peak traffic loads, considering burst characteristics.

CCNA Exam Relevance

On the CCNA 200-301 exam, this scenario appears in troubleshooting questions where you must identify the cause of output drops. Expect multiple-choice questions asking which command to use to check for drops, or drag-and-drop steps to resolve the issue. Key fact: Output drops indicate the transmit ring is full, and the default FIFO queue has a limited size (40 packets).

Exam Tips

1.

Remember that 'show interfaces' includes 'Total output drops' and 'Output queue: size/max' — these are the primary indicators of interface congestion.

2.

The default queue size on Cisco switches is often 40 packets; increasing it or applying QoS can reduce drops.

3.

Know that 'show queueing interface' shows the current queue depth and drops, and 'show policy-map interface' confirms if QoS is applied.

Commands Used in This Scenario

Test Your CCNA Knowledge

Practice with scenario-based questions to prepare for the CCNA 200-301 exam.

Practice CCNA Questions