Wireshark is a packet analyzer program, formerly known as Ethereal, that supports multiple protocols and presents information in a text-based user interface.

The ability to capture and analyze traffic provides data on network activity. Prior to Cisco IOS Release XE 3.3.0(SE), only two features addressed this need: SPAN and debug platform packet. Both have limitations. SPAN is ideal for capturing packets, but can only deliver them by forwarding them to some specified local or remote destination; it provides no local display or analysis support.

So the need exists for a traffic capture and analysis mechanism that is applicable to both hardware and software forwarded traffic and that provides strong packet capture, display, and analysis support, preferably using a well known interface.

Wireshark dumps packets to a file using a well known format called .pcap, and is applied or enabled on individual interfaces. You specify an interface in EXEC mode along with the filter and other parameters. The Wireshark application is applied only when you enter a start command, and is removed only when Wireshark stops capturing packets either automatically or manually.

Note	The current version of Wireshark installed on the switch is 1.10.8.

A capture point is the central policy definition of the Wireshark feature. The capture point describes all of the characteristics associated with a given instance of Wireshark: which packets to capture, where to capture them from, what to do with the captured packets, and when to stop. Capture points can be modified after creation, and do not become active until explicitly activated with a start command. This process is termed activating the capture point or starting the capture point. Capture points are identified by name and can also be manually or automatically deactivated or stopped.

Multiple capture points can be defined, but only one can be active at a time. You need to stop one before you can start the other.

In case of stacked systems, the capture point is activated on the active member. A switchover will terminate any active packet capture session and it will have to be restarted.

An attachment point is a point in the logical packet process path associated with a capture point. An attachment point is an attribute of the capture point. Packets that impact an attachment point are tested against capture point filters; packets that match are copied and sent to the associated Wireshark instance of the capture point. A specific capture point can be associated with multiple attachment points, with limits on mixing attachment points of different types. Some restrictions apply when you specify attachment points of different types. Attachment points are directional (input or output or both) with the exception of the Layer 2 VLAN attachment point, which is always bidirectional.

In case of stacked systems, the attachment points on all stack members are valid. EPC captures the packets from all the defined attachment points. However these packets are processed only on the active member.

Filters are attributes of a capture point that identify and limit the subset of traffic traveling through the attachment point of a capture point, which is copied and passed to Wireshark. To be displayed by Wireshark, a packet must pass through an attachment point, as well as all of the filters associated with the capture point.

A capture point has the following types of filters:

• Core system filter—The core system filter is applied by hardware, and its match criteria is limited by hardware. This filter determines whether hardware-forwarded traffic is copied to software for Wireshark purposes.

• Display filter—The display filter is applied by Wireshark. Packets that fail the display filter are not displayed.

Core System Filter

You can specify core system filter match criteria by using the class map or ACL, or explicitly by using the CLI.

Note	When specifying CAPWAP as an attachment point, the core system filter is not used.

In some installations, you need to obtain authorization to modify the configuration, which can lead to extended delays if the approval process is lengthy. This can limit the ability of network administrators to monitor and analyze traffic. To address this situation, Wireshark supports explicit specification of core system filter match criteria from the EXEC mode CLI. The disadvantage is that the match criteria that you can specify is a limited subset of what class map supports, such as MAC, IP source and destination addresses, ether-type, IP protocol, and TCP/UDP source and destination ports.

If you prefer to use configuration mode, you can define ACLs or have class maps refer capture points to them. Explicit and ACL-based match criteria are used internally to construct class maps and policy maps.

Note The ACL and class map configuration are part of the system and not aspects of the Wireshark feature.

Wireshark can be invoked on live traffic or on a previously existing .pcap file. When invoked on live traffic, it can perform four types of actions on packets that pass its display filters:

• Captures to buffer in memory to decode and analyze and store

• Stores to a .pcap file

• Decodes and displays

• Stores and displays

When invoked on a .pcap file only, only the decode and display action is applicable.

Packets can be stored in the capture buffer in memory for subsequent decode, analysis, or storage to a .pcap file.

The capture buffer can be in linear or circular mode. In linear mode, new packets are discarded when the buffer is full. In circular mode, if the buffer is full, the oldest packets are discarded to accommodate the new packets. Although the buffer can also be cleared when needed, this mode is mainly used for debugging network traffic. However, it is not possible to only clear the contents of the buffer alone without deleting it. Stop the current captures and restart the capture again for this to take effect.

Note	If you have more than one capture that is storing packets in a buffer, clear the buffer before starting a new capture to avoid memory loss.

Note

When WireShark is used on switches in a stack, packet captures can be stored only on flash or USB flash devices connected to the active switch. For example, if flash1 is connected to the active switch, and flash2 is connected to the secondary switch, only flash1 can be used to store packet captures. Attempts to store packet captures on devices other than flash or USB flash devices connected to the active switch will probably result in errors.

Wireshark can store captured packets to a .pcap file. The capture file can be located on the following storage devices:

• on-board flash storage (flash:)

• USB drive (usbflash0:)

Note	Attempts to store packet captures on unsupported devices or devices not connected to the active switch will probably result in errors.

When configuring a Wireshark capture point, you can associate a filename. When the capture point is activated, Wireshark creates a file with the specified name and writes packets to it. If the file already exists at the time of creation of the capture point, Wireshark queries you as to whether the file can be overwritten. If the file already exists at the time of activating the capture point, Wireshark will overwrite the existing file. Only one capture point may be associated with a given filename.

If the destination of the Wireshark writing process is full, Wireshark fails with partial data in the file. You must ensure that there is sufficient space in the file system before you start the capture session. With Cisco IOS Release IOS XE 3.3.0(SE), the file system full status is not detected for some storage devices.

You can reduce the required storage space by retaining only a segment, instead of the entire packet. Typically, you do not require details beyond the first 64 or 128 bytes. The default behavior is to store the entire packet.

To avoid possible packet drops when processing and writing to the file system, Wireshark can optionally use a memory buffer to temporarily hold packets as they arrive. Memory buffer size can be specified when the capture point is associated with a .pcap file.

Wireshark can decode and display packets to the console. This functionality is possible for capture points applied to live traffic and for capture points applied to a previously existing .pcap file.

Note	Decoding and displaying packets may be CPU intensive.

Wireshark can decode and display packet details for a wide variety of packet formats. The details are displayed by entering the monitor capture name start command with one of the following keyword options, which place you into a display and decode mode:

• brief—Displays one line per packet (the default).

• detailed—Decodes and displays all the fields of all the packets whose protocols are supported. Detailed modes require more CPU than the other two modes.

• (hexadecimal) dump—Displays one line per packet as a hexadecimal dump of the packet data and the printable characters of each packet.

When you enter the capture command with the decode and display option, the Wireshark output is returned to Cisco IOS and displayed on the console unchanged.

Functionally, this mode is a combination of the previous two modes. Wireshark stores packets in the specified .pcap file and decodes and displays them to the console. Only the core filters are applicable here.

After a Wireshark capture point has been defined with its attachment points, filters, actions, and other options, it must be activated. Until the capture point is activated, it does not actually capture packets.

Before a capture point is activated, some functional checks are performed. A capture point cannot be activated if it has neither a core system filter nor attachment points defined. Attempting to activate a capture point that does not meet these requirements generates an error.*

Note	*When performing a wireless capture with a CAPWAP tunneling interface, the core system filter is not required and cannot be used.

The display filters are specified as needed.

After Wireshark capture points are activated, they can be deactivated in multiple ways. A capture point that is storing only packets to a .pcap file can be halted manually or configured with time or packet limits, after which the capture point halts automatically.

When a Wireshark capture point is activated, a fixed rate policer is applied automatically in the hardware so that the CPU is not flooded with Wireshark-directed packets. The disadvantage of the rate policer is that you cannot capture contiguous packets beyond the established rate even if more resources are available.

The set packet capture rate is 1000 packets per sec (pps). The 1000 pps limit is applied to the sum of all attachment points. For example, if we have a capture session with 3 attachment points, the rates of all 3 attachment points added together is policed to 1000 pps.

Note	Policer is not supported for control-plane packet capture. When activating control-plane capture points, you need to be extra cautious, so that it does not flood the CPU.

This section describes how Wireshark features function in the environment:

• If port security and Wireshark are applied on an ingress capture, a packet that is dropped by port security will still be captured by Wireshark. If port security is applied on an ingress capture, and Wireshark is applied on an egress capture, a packet that is dropped by port security will not be captured by Wireshark.

• Packets dropped by Dynamic ARP Inspection (DAI) are not captured by Wireshark.

• If a port that is in STP blocked state is used as an attachment point and the core filter is matched, Wireshark will capture the packets that come into the port, even though the packets will be dropped by the switch.

• Classification-based security features—Packets that are dropped by input classification-based security features (such as ACLs and IPSG) are not caught by Wireshark capture points that are connected to attachment points at the same layer. In contrast, packets that are dropped by output classification-based security features are caught by Wireshark capture points that are connected to attachment points at the same layer. The logical model is that the Wireshark attachment point occurs after the security feature lookup on the input side, and symmetrically before the security feature lookup on the output side.

  On ingress, a packet goes through a Layer 2 port, a VLAN, and a Layer 3 port/SVI. On egress, the packet goes through a Layer 3 port/SVI, a VLAN, and a Layer 2 port. If the attachment point is before the point where the packet is dropped, Wireshark will capture the packet. Otherwise, Wireshark will not capture the packet. For example, Wireshark capture policies connected to Layer 2 attachment points in the input direction capture packets dropped by Layer 3 classification-based security features. Symmetrically, Wireshark capture policies attached to Layer 3 attachment points in the output direction capture packets dropped by Layer 2 classification-based security features.

• Routed ports and switch virtual interfaces (SVIs)—Wireshark cannot capture the output of an SVI because the packets that go out of an SVI's output are generated by CPU. To capture these packets, include the control plane as an attachment point.

• VLANs—Starting with Cisco IOS Release 16.1, when a VLAN is used as a Wireshark attachment point, packet capture is supported on L2 and L3 in both input and output directions.

• Redirection features—In the input direction, features traffic redirected by Layer 3 (such as PBR and WCCP) are logically later than Layer 3 Wireshark attachment points. Wireshark captures these packets even though they might later be redirected out another Layer 3 interface. Symmetrically, output features redirected by Layer 3 (such as egress WCCP) are logically prior to Layer 3 Wireshark attachment points, and Wireshark will not capture them.

• SPAN—Wireshark cannot capture packets on interface configured as a SPAN destination.

• SPAN—Wireshark is able to capture packets on interfaces configured as a SPAN source in the ingress direction, and may be available for egress direction too.

• You can capture packets from a maximum of 1000 VLANs at a time, if no ACLs are applied. If ACLs are applied, the hardware will have less space for Wireshark to use. As a result, the maximum number of VLANs than can be used for packet capture at a time will be lower. Using more than 1000 VLANs tunnels at a time or extensive ACLs might have unpredictable results. For example, mobility may go down.

  Note	Capturing an excessive number of attachment points at the same time is strongly discouraged because it may cause excessive CPU utilization and unpredictable hardware behavior.

• During Wireshark packet capture, hardware forwarding happens concurrently.

• Before starting a Wireshark capture process, ensure that CPU usage is moderate and that sufficient memory (at least 200 MB) is available.

• If you plan to store packets to a storage file, ensure that sufficient space is available before beginning a Wireshark capture process.

• The CPU usage during Wireshark capture depends on how many packets match the specified conditions and on the intended actions for the matched packets (store, decode and display, or both).

• Where possible, keep the capture to the minimum (limit by packets, duration) to avoid high CPU usage and other undesirable conditions.

• Because packet forwarding typically occurs in hardware, packets are not copied to the CPU for software processing. For Wireshark packet capture, packets are copied and delivered to the CPU, which causes an increase in CPU usage.

  To avoid high CPU usage, do the following:

  • Attach only relevant ports.

  • Use a class map, and secondarily, an access list to express match conditions. If neither is viable, use an explicit, in-line filter.

  • Adhere closely to the filter rules. Restrict the traffic type (such as, IPv4 only) with a restrictive, rather than relaxed ACL, which elicits unwanted traffic.

• Always limit packet capture to either a shorter duration or a smaller packet number. The parameters of the capture command enable you to specify the following:

  • Capture duration

  • Number of packets captured

  • File size

  • Packet segment size

• Run a capture session without limits if you know that very little traffic matches the core filter.

• You might experience high CPU (or memory) usage if:

  • You leave a capture session enabled and unattended for a long period of time, resulting in unanticipated bursts of traffic.

  • You launch a capture session with ring files or capture buffer and leave it unattended for a long time, resulting in performance or system health issues.

• During a capture session, watch for high CPU usage and memory consumption due to Wireshark that may impact performance or health. If these situations arise, stop the Wireshark session immediately.

• Avoid decoding and displaying packets from a .pcap file for a large file. Instead, transfer the .pcap file to a PC and run Wireshark on the PC.

• You can define up to eight Wireshark instances. An active show command that decodes and displays packets from a .pcap file or capture buffer counts as one instance. However, only one of the instances can be active.

• Whenever an ACL that is associated with a running capture is modified, you must restart the capture for the ACL modifications to take effect. If you do not restart the capture, it will continue to use the original ACL as if it had not been modified.

• To avoid packet loss, consider the following:

  • Use store-only (when you do not specify the display option) while capturing live packets rather than decode and display, which is an CPU-intensive operation (especially in detailed mode).

  • If you have more than one capture that is storing packets in a buffer, clear the buffer before starting a new capture to avoid memory loss.

  • If you use the default buffer size and see that you are losing packets, you can increase the buffer size to avoid losing packets.

  • Writing to flash disk is a CPU-intensive operation, so if the capture rate is insufficient, you may want to use a buffer capture.

  • The Wireshark capture session always operates in streaming mode at the rate of 1000 pps.

• The streaming capture mode rate is 1000 pps.

• If you want to decode and display live packets in the console window, ensure that the Wireshark session is bounded by a short capture duration.

Warning

A Wireshark session with either a longer duration limit or no capture duration (using a terminal with no auto-more support using the term len 0 command) may make the console or terminal unusable.

• When using Wireshark to capture live traffic that leads to high CPU, usage, consider applying a QoS policy temporarily to limit the actual traffic until the capture process concludes.

• All Wireshark-related commands are in EXEC mode; no configuration commands exist for Wireshark.

  If you need to use access list or class-map in the Wireshark CLI, you must define an access list and class map with configuration commands.

• No specific order applies when defining a capture point; you can define capture point parameters in any order, provided that CLI allows this. The Wireshark CLI allows as many parameters as possible on a single line. This limits the number of commands required to define a capture point.

• All parameters except attachment points take a single value. Generally, you can replace the value with a new one by reentering the command. After user confirmation, the system accepts the new value and overrides the older one. A no form of the command is unnecessary to provide a new value, but it is necessary to remove a parameter.

• Wireshark allows you to specify one or more attachment points. To add more than one attachment point, reenter the command with the new attachment point. To remove an attachment point, use the no form of the command. You can specify an interface range as an attachment point. For example, enter monitor capture mycap interface GigabitEthernet1/0/1 in where interface GigabitEthernet1/0/1 is an attachment point.

  If you also need to attach interface GigabitEthernet1/0/2, specify it in another line as follows:

  monitor capture mycap interface GigabitEthernet1/0/2 in

• You cannot make changes to a capture point when the capture is active.

• The action you want to perform determines which parameters are mandatory. The Wireshark CLI allows you to specify or modify any parameter prior to entering the start command. When you enter the start command, Wireshark will start only after determining that all mandatory parameters have been provided.

• If the file already exists at the time of creation of the capture point, Wireshark queries you as to whether the file can be overwritten. If the file already exists at the time of activating the capture point, Wireshark will overwrite the existing file.

• The core filter can be an explicit filter, access list, or class map. Specifying a newer filter of these types replaces the existing one.

  Note	A core filter is required except when using a CAPWAP tunnel interface as a capture point attachment point.

• You can terminate a Wireshark session with an explicit stop command or by entering q in automore mode. The session could terminate itself automatically when a stop condition such as duration or packet capture limit is met, or if an internal error occurs, or resource is full (specifically if disk is full in file mode).

• Dropped packets will not be shown at the end of the capture. However, only the count of dropped, oversized packets will be displayed.

The table below shows the default Wireshark configuration.

Embedded Packet Capture (EPC) provides an embedded systems management facility that helps in tracing and troubleshooting packets. This feature allows network administrators to capture data packets flowing through, to, and from a Cisco device. The network administrator may define the capture buffer size and type (circular, or linear) and the maximum number of bytes of each packet to capture. The packet capture rate can be throttled using further administrative controls. For example, options allow for filtering the packets to be captured using an Access Control List and, optionally, further defined by specifying a maximum packet capture rate or by specifying a sampling interval.

• Ability to capture IPv4 and IPv6 packets in the device, and also capture non-IP packets with MAC filter or match any MAC address.

• Extensible infrastructure for enabling packet capture points. A capture point is a traffic transit point where a packet is captured and associated with a buffer.

• Facility to export the packet capture in packet capture file (PCAP) format suitable for analysis using any external tool.

• Methods to decode data packets captured with varying degrees of detail.

Packet data capture is the capture of data packets that are then stored in a buffer. You can define packet data captures by providing unique names and parameters.

You can perform the following actions on the capture:

• Activate captures at any interface.

• Apply access control lists (ACLs) or class maps to capture points.

  Note	Network Based Application Recognition (NBAR) and MAC-­style class map is not supported.

• Destroy captures.

• Specify buffer storage parameters such as size and type. The size ranges from 1 MB to 100 MB. The default buffer is linear;; the other option for the buffer is circular.

• Specify match criteria that includes information about the protocol, IP address or port address.