VMware ESXi and vSphere Cluster Management

Cisco IOS Extended Ping: Testing Reachability, Source Paths, MTU, and Packet Behavior

Learn how to use Cisco IOS extended ping to select source addresses, test return routes, investigate MTU and DF behavior, inspect loss and latency, and validate packet handling.

What Cisco IOS Extended Ping Tests

Ping uses an ICMP Echo Request and waits for an ICMP Echo Reply. A reply shows that a packet reached the destination and that return traffic came back under the conditions of the test.

Ordinary ping uses mostly default values and specifies the destination on the command line:

Router# ping 10.0.0.2

Extended ping is Cisco IOS's interactive ping mode. Start it from privileged EXEC mode by entering ping without a destination:

Router# ping

The dialog lets you change packet and test characteristics, such as the source address, repeat count, datagram size, timeout, service marking, Don't Fragment behavior, and data pattern. Prompt order and available options vary by IOS release, platform, protocol, and IPv4 or IPv6 mode.

Basic Extended-Ping Workflow

  1. Enter privileged EXEC mode, normally identified by a prompt ending in #.
  2. Enter ping without a destination.
  3. Select the protocol. ip is normally the relevant choice for IPv4 troubleshooting.
  4. Enter the destination IP address or a resolvable hostname.
  5. Accept or change the repeat count, datagram size, and timeout.
  6. Choose whether to use extended commands.
  7. If enabled, answer the additional prompts, such as source address, service value, DF behavior, data validation, or packet-size sweep.
  8. Review the success indicators, success rate, and round-trip-time summary.

Use context-sensitive help and the prompts displayed by the device rather than assuming that every IOS version presents the same questions.

Protocol Selection

IP is the standard selection for modern IPv4-network troubleshooting. Some IOS versions or platforms historically present additional protocol choices. Those legacy choices are platform-dependent and generally do not apply to current IPv4 or IPv6 networks.

Ordinary Ping Versus Interactive Extended Ping

CapabilityOrdinary pingInteractive extended pingDiagnostic value
Destination testYes, supplied on the command lineYes, supplied during promptsTests basic ICMP reachability
Source selectionUsually uses the normal routed sourceCan select an interface or IPv4 addressTests return routing to a particular source subnet
Probe countUses command defaults or limited syntaxCan change repeat countHelps reveal intermittent loss
Payload sizeUses a defaultCan change datagram sizeSupports MTU investigation
TimeoutUses a defaultCan change the wait timeUseful on slow or high-delay paths
DF and MTU testingNot normally exposed interactivelyAvailable where supportedTests fragmentation and path-MTU limits
QoS markingUsually uses the default markingCan expose a Type of Service promptHelps verify classification policies
Data-pattern testingUses ordinary defaultsCan expose pattern and validation controlsSupports specialized link troubleshooting

Core Ping Parameters

IOS prompt or parameterTypical defaultWhat it changesWhen to modify itOperational cautions
ProtocolipSelects the protocol used by the testUse IP for normal IPv4 diagnosticsChoices differ by platform and release
Target addressNoneDestination IP address or resolvable hostnameTest a router interface, host, or service-side addressName resolution can introduce a separate DNS dependency
Repeat countOften 5, but verify locallyNumber of Echo Requests sentIncrease it when investigating intermittent lossMore probes create more traffic and take longer
Datagram sizeOften 100, but verify locallyConfigured test data size presented by IOSTest fragmentation, MTU limits, or link behaviorDo not confuse the configured size with total on-wire size
TimeoutOften 2 seconds, but verify locallyMaximum wait for each replyIncrease it for high-delay pathsA long timeout can make a large test slow to complete
Extended commandsUsually noExposes advanced promptsChoose yes for source, DF, QoS, or specialized testsPrompt wording and options vary
Source address or interfaceRouting-selected sourceSets the source field of the Echo RequestTest reachability from a particular interface or subnetThe destination must have a return route to that source
Type of ServicePlatform defaultSets a service-field value where supportedCheck QoS classification or policy behaviorValues and wording differ; confirm the platform's interpretation
DF bitUsually disabledPrevents IPv4 fragmentation when enabledFind a path-MTU limitControl messages may be filtered, obscuring the reason for failure
Validate reply dataUsually disabledChecks whether returned data matches sent dataInvestigate suspected corruptionSupported behavior is platform-dependent
Data patternImplementation-specific hexadecimal defaultChanges the contents of the test dataInvestigate framing, CRC, clocking, or corruption symptomsVerify the actual default on the device
IP header optionsUsually noneMay request source routing, record route, timestamps, or verbose detailsUse only for specialized diagnosticsOptions may be unsupported, filtered, deprecated, or unsafe in production
Size sweepUsually disabledSends probes over a range of packet sizesFind where loss or fragmentation-related failure beginsPrompt behavior is IOS- and platform-dependent

Launching a Baseline Extended Ping

First establish a simple comparison result:

Router# ping 10.0.0.2

Then launch the interactive form. The following values are illustrative; actual defaults must be read from the device:

Router# ping
Protocol [ip]: ip
Target IP address: 10.0.0.2
Repeat count [5]: 5
Datagram size [100]: 100
Timeout in seconds [2]: 2
Extended commands [n]: n

Press Enter to accept a displayed default. A baseline test normally keeps the default packet size, timeout, source behavior, and non-extended settings. This makes later tests easier to compare.

Selecting a Source Interface or Address

The source address is the IPv4 address placed in the source field of the Echo Request. A source interface is an interface selected to supply that address. The egress interface is the outbound interface chosen by routing for the destination.

By default, IOS generally selects an address associated with the egress interface. This is convenient for ordinary reachability testing, but it may not test every local interface or subnet.

With extended ping, select a source address or, on supported platforms, an interface:

Router# ping
Protocol [ip]: ip
Target IP address: 10.0.0.2
Repeat count [5]: 5
Datagram size [100]: 100
Timeout in seconds [2]: 2
Extended commands [n]: y
Source address or interface: 192.168.5.1

Choosing an alternate source turns the test into a routing test in both directions. The local router must have a usable forward route to the destination, and the destination or an upstream router must have a return path to the selected source address. A return path is the route used by replies to travel back to the ping source.

Two-Router Source-Address Example

Consider this topology:

  • R2 Fa0/1: 10.0.0.1/24
  • R1 Fa0/0: 10.0.0.2/24
  • R2 Fa0/0: 192.168.5.1/24
  • R1 needs a route to 192.168.5.0/24 through R2.

A normal ping from R2 to 10.0.0.2 will likely use 10.0.0.1 as its source because that is the address on the interface toward R1. Now repeat the test while selecting 192.168.5.1 or R2 Fa0/0:

R2# ping
Protocol [ip]: ip
Target IP address: 10.0.0.2
Repeat count [5]: 5
Datagram size [100]: 100
Timeout in seconds [2]: 2
Extended commands [n]: y
Source address or interface: 192.168.5.1

If replies arrive, R1 or the routed network beyond it can return traffic toward 192.168.5.0/24. If the default test succeeds but this test times out, inspect the route from R1 toward 192.168.5.0/24. Also check ACLs, NAT, VRF boundaries, interface state, and source-address correctness before concluding that the forward link is broken.

Reading IOS Ping Output

IOS commonly displays one symbol per probe and then summarizes the result. A success symbol such as ! indicates a reply. A timeout symbol such as . indicates that no reply was received before the timeout. An unreachable indicator such as U indicates that an ICMP Unreachable message was received. Exact symbols can vary by protocol and platform.

The success rate compares replies received with probes sent. For example, 80 percent means that four of five probes received replies. The RTT summary reports delay, often as minimum, average, and maximum values. RTT can reveal delay and variation, but ping alone does not prove application performance.

Observed resultLikely meaningInitial checksWhat the result does not prove
All replies succeedICMP reached the destination and replies returnedCheck RTT, route, ACLs, and the intended sourceDoes not validate TCP or UDP ports, DNS, or application health
All probes time outNo reply arrived before the timeoutCheck source and destination interfaces, routes, ACLs, firewall policy, and ICMP filteringDoes not prove that all IP or application traffic fails
Partial successLoss, congestion, unstable paths, rate limiting, or inconsistent forwarding may existIncrease repeat count; inspect drops, errors, CPU, QoS, and ECMP pathsDoes not identify the failing layer by itself
Unreachable responseA device reported that delivery or the route was unavailableIdentify the responding hop and inspect routing and interface stateDoes not always identify the original cause of the unreachable condition
High or variable RTTDelay, queuing, congestion, or a slow control plane may be presentCompare repeated runs and inspect utilization and QoSDoes not directly measure application latency
Failure only with an alternate sourceThe selected source's return path or policy is brokenCheck the remote route to the source subnet, ACLs, NAT, VRF, and interface stateDoes not necessarily indicate a forward-path failure
Failure only at larger DF-set sizesA path MTU limit, tunnel overhead, or blocked fragmentation-needed message may existReduce size, calculate packet overhead, and inspect MTU settingsDoes not by itself identify which hop has the lower MTU

Type of Service, DSCP, and QoS Testing

Type of Service (ToS) is a historical description of service-handling information in the IP header. Modern networks commonly interpret the service field using DSCP, or Differentiated Services Code Point, for QoS marking.

Where IOS supports the relevant prompt, marking an extended-ping packet can help test whether a classifier recognizes the expected value. Inspect policy counters or capture the packet to verify classification. A successful Echo Reply only proves that the packet was returned; it does not prove that every intended QoS action, queue, or priority treatment was applied.

Prompt wording, numeric values, and supported meanings differ across IOS releases. Use the device's displayed options and QoS documentation.

Don't Fragment and Path MTU Testing

The IPv4 Don't Fragment (DF) bit tells a router not to fragment the packet. If a DF-set packet is larger than an interface or path can carry, the packet cannot be fragmented and should fail at the limiting point. A router may send an ICMP fragmentation-needed message, but filtering that message can make the symptom look like a timeout.

MTU, or Maximum Transmission Unit, is the largest packet size an interface can carry without required fragmentation. The path MTU is the smallest effective MTU across the entire route. Tunnels, VPNs, WAN encapsulation, and other overhead can reduce it.

  1. Begin with a known working, conservative size.
  2. Enable extended commands and set DF according to the platform's prompt.
  3. Increase the datagram size in controlled steps.
  4. Record the largest successful size and the first failing size.
  5. Inspect interface, tunnel, and encapsulation MTUs and check filtering of ICMP control messages.

The size shown by an IOS prompt may represent the ICMP data or datagram test size rather than the complete on-wire IP packet. Total IPv4 packet size includes the configured data plus the IP header and ICMP header. For a basic IPv4 Echo packet with no IP options, the common calculation is:

total IP packet size = ICMP data size + 20-byte IPv4 header + 8-byte ICMP header

Therefore, do not compare a prompt value directly with an interface MTU without accounting for header overhead and any tunnel or WAN encapsulation.

Packet-Size Sweeps

A size sweep sends multiple probes over a specified range of sizes. It can show where loss or fragmentation-related failure begins and is useful for estimating a limiting path MTU. Sweep prompts and syntax are IOS- and platform-dependent, so use only the parameters offered by the device.

Payload Validation and Data Patterns

Payload validation checks whether returned Echo data matches the data that was sent. If the platform supports this option, it can help detect data alteration rather than merely checking that an Echo Reply arrived.

A data pattern is a chosen bit pattern placed in the test data. Nondefault patterns can help investigate suspected corruption, framing errors, or clocking problems, especially on serial or legacy WAN links. The default hexadecimal pattern is an implementation default; verify it on the actual device instead of assuming it is universal.

Ping results are not sufficient to confirm a physical or data-link fault. Correlate the test with show interfaces, physical diagnostics, CRC and framing counters, clocking information, circuit-provider data, and error logs.

IP Header Options

Some IOS versions expose specialized IP options:

  • Loose source route: requests that traffic pass through specified routers, while allowing other hops between them.
  • Strict source route: requests that traffic follow specified routers directly and in the specified order.
  • Record route: asks routers to record addresses encountered along the path.
  • Timestamp: requests timing information from participating devices.
  • Verbose-related options: provide additional diagnostic output where supported.

These features are not routine connectivity tests. They may be unsupported, filtered by security devices, deprecated, or unsuitable for production traffic. Use them only when the platform and network policy explicitly support the test.

Practical Diagnostic Examples

Investigate Intermittent Loss

Increase the repeat count and use a timeout appropriate for the path. Compare the success percentage and RTT range across multiple runs. Then correlate the result with interface drops and errors, congestion, CPU utilization, QoS policing, and ICMP rate limiting. Partial success can also result from ECMP paths with different reachability or MTU characteristics.

Verify a QoS Marking

Set a supported ToS or DSCP-related value in the extended dialog. Inspect policy counters or capture the packet to confirm the mark and classification. Do not infer that a reply proves all QoS actions were applied.

Investigate Serial-Link Corruption

Select a supported alternate data pattern and enable reply-data validation if available. Send enough probes to make the symptom reproducible. Confirm the finding with interface CRC, framing, input-error, and clocking counters plus physical-circuit diagnostics.

Troubleshooting Checklist

Alternate Source Fails but Default Ping Works

  • Verify that the selected source address belongs to an up/up interface and is correctly configured.
  • Check the local route toward the destination.
  • Check the remote or upstream route toward the selected source subnet.
  • Review ACLs, firewall rules, NAT, VRF boundaries, and control-plane policies in both directions.
  • Retest after correcting the return route or policy.

Small Packets Work but Large DF-Set Packets Fail

  • Reduce the size until DF-set probes succeed.
  • Calculate total packet size, including IP and ICMP headers.
  • Inspect interface and tunnel MTUs and encapsulation overhead.
  • Check whether ICMP fragmentation-needed messages are filtered.

Useful Verification Commands

Router# show ip route
Router# show ip interface brief
Router# show interfaces
Router# show access-lists
Router# traceroute 10.0.0.2

Use the routing table to inspect the destination route and the expected return route. Use interface output to check administrative state, line protocol state, addresses, errors, and drops. Traceroute investigates the forward path, but its source-selection syntax and capabilities are platform-dependent.

Limits of Ping

A successful ping verifies only that ICMP traffic was returned under the tested conditions. It does not independently validate TCP or UDP ports, DNS name resolution, application health, authentication, or every possible path.

Ping can fail while an application works because a firewall or ACL filters ICMP, a device rate-limits or deprioritizes ICMP, or the application uses a permitted TCP or UDP flow. When this occurs, review ICMP policy and test the actual application port with an appropriate tool.

If an option described in a guide is absent, the IOS version, feature set, platform, protocol, or address family may differ. Use CLI help and device documentation, and record the software and platform when comparing behavior.

Continue with the focused guide: Use Extended Ping.