Nexus Platform Tools
Nexus switches are among the most powerful data center switches in the industry. This is partly because of the CPU and memory available in the switch, but also because of the wide range of integrated tools that the NX-OS offers. These tools provide the capability to capture packets at different ASIC levels within the switch and help verify both hardware programming and the action taken by the hardware or the software on the packet under investigation. Some of these tools include the following:
Ethanalyzer
Embedded Logic Analyzer Module (ELAM)
Packet Tracer
These tools are capable of performing packet capture for the traffic destined for the CPU or transit hardware-switched traffic. They are helpful in understanding the stages the packet goes through in a switch, which helps narrow down the issue very quickly. The main benefit of these features is that they do not require time to set up an external sniffing device.
Ethanalyzer
Ethanalyzer is an NX-OS implementation of TShark, a terminal version of Wireshark. TShark uses the libpcap library, which gives Ethanalyzer the capability to capture and decode packets. It can capture inband and management traffic on all Nexus platforms. Ethanalyzer provides the users with the following capabilities:
Capture packets sent and received by the switch Supervisor CPU
Define the number of packets to be captured
Define the length of the packets to be captured
Display packets with very detailed protocol information or a one-line summary
Open and save captured packet data
Filter packets capture on many criteria (capture filter)
Filter packets to be displayed on many criteria (display filter)
Decode the internal header of control packet
Avoid the requirement of using an external sniffing device to capture the traffic
Ethanalyzer does not allow hardware-switched traffic to be captured between data ports of the switch. For this type of packet capture, SPAN or ELAM is used. When the interfaces are configured with ACLs with ACEs configured with the log option, the hardware-switched flows gets punted to the CPU and thus are captured using Ethanalyzer. However, this should not be tried in production because the packets could get dropped as a result of CoPP policies or the excessive traffic punted to the CPU could impact other services on the device.
Ethanalyzer is configured in three simple steps:
Step 1. Define capture interface.
Step 2. Define Filters: Set the capture filter or display filter.
Step 3. Define the stop criteria.
There are three kinds of capture interfaces:
Mgmt: Captures traffic on the Mgmt0 interface of the switch
Inbound-hi: Captures high-priority control packets on the inband, such as Spanning Tree Protocol (STP), Link Aggregation Control Protocol (LACP), Cisco Discovery Protocol (CDP), Data Center Bridging Exchange (DCBX), Fiber Channel, and Fiber Channel over Ethernet (FCOE)
Inbound-low: Captures low-priority control packets on the inband, such as Internet Group Management Protocol (IGMP), Transmission Control Protocol (TCP), User Datagram Protocol (UDP), Internet Protocol (IP), and Address Resolution Protocol (ARP) traffic
The next step is to set the filters. With a working knowledge of Wireshark, configuring filters for Ethanalyzer is fairly simple. Two kinds of filters can be set up for configuring Ethanalyzer: capture filter and display filter. As the name suggests, when a capture filter is set, only frames that match the filter are captured. The display filter is used to display the packets that match the filter from the captured set of packets. That means Ethanalyzer captures other frames that do not match the display filter but are not displayed in the output. By default, Ethanalyzer supports capturing up to 10 frames and then stops automatically. This value is changed by setting the limit-captured-frames option, where 0 means no limit.
Ethanalyzer is part of the software running on the supervisor, so it is important to understand its effects on the supervisor’s CPU. Normally, Ethanalyzer does not have much impact, but sometimes it can increase the CPU utilization up to 5%. Utilization can be reduced by 1% to 2% by saving the capture data in a file using the write option with Ethanalyzer to save the capture in a file.
To start a packet capture with Ethanalyzer, use the command ethanalyzer local interface [inbound-hi | inbound-lo | mgmt] options, with the following options:
Autostop: Capture autostop condition
capture-filter: Filter on Ethanalyzer capture
capture-ring-buffer: Capture ring buffer option
decode-internal: Include internal system header decoding
detail: Display detailed protocol information
display-filter: Display filter on frames captured
limit-captured-frames: Indicates the maximum number of frames to be captured
limit-frame-size: Capture only a subset of a frame
write: Identifies the filename to save capture to
While using Ethanalyzer, specifying the filters is easier for someone who is familiar with Wireshark filters. The syntax for both the capture filter and the display filter is different. Table 2-1 lists some of the common filters and their syntax with the capture-filter and display-filter options.
Table 2-1 Ethanalyzer Capture and Display Filters
|
Capture Filter |
Display Filter |
Operators |
|
And - && Or - || Equal - == Not equal - != |
VLAN |
vlan vlan-id |
vlan.id==vlan-id |
Layer 2 |
ether host 00:AA:BB:CC:DD:EE ether dst 00:AA:BB:CC:DD:EE ether src 00:AA:BB:CC:DD:EE ether broadcast ether multicast ether proto protocol |
eth.addr==00:AA:BB:CC:DD:EE eth.src==00:AA:BB:CC:DD:EE eth.dst==00:AA:BB:CC:DD:EE Match first 2 bytes: eth.src[0:1]==00:AA Filter on manufacturer: eth.src[0:2]==vendor-mac-addr e.g., Cisco: eth.src[0:2]==00.00.0c eth.addr contains aa:bb:cc |
Layer 3 |
ip (filters out lower-level protocols such as ARP and STP) host 192.168.1.1 dst host 192.168.1.1 src host 192.168.1.1 net 192.168.1.0/24 net 192.168.1.0 netmask 24 src net 192.168.1.0/24 dst net 192.168.1.0/24 ip broadcast ip multicast not broadcast not multicast icmp udp tcp ip proto 6 (udp) ip proto 17 (tcp) ip proto 1 (icmp) Packet length: less length greater length |
IP address: ip.addr==192.168.1.1 Source IP: ip.src==192.168.1.1 Dest IP: ip.dst==192.168.10.1 Subnet: ip.addr==192.168.1.0/24 Fragmentation: Filter on DF bit set (0 = may fragment) ip.flags.df==1 TCP Sequence: tcp.seq==TCP-Seq-Num |
Layer 4 |
udp port 53 udp dst port 53 udp src port 53 tcp port 179 tcp portrange 2000-2100 |
tcp.port==53 udp.port==53 |
FabricPath |
proto 0x8903 |
Dest HMAC/MC destination: cfp.d_hmac==mac cfp.d_hmac_mc==mac EID/FTAG/IG Bit: cfp.eid== cfp.ftag== cfp.ig== Source LID/OOO/DL Bit/Source HMAC: cfp.lid== cfp.ooodl== cfp.s_hmac== Subswitch ID/Switch ID/TTL: cfp.sswid== cfp.swid== cfp.ttl== |
ICMP |
icmp |
icmp==icmp-type ICMP-Types: icmp-echoreply icmp-unreach icmp-sourcequench icmp-redirect icmp-echo icmp-routeradvert icmp-routersolicit icmp-timxceed icmp-paramprob icmp-tstamp icmp-tstampreply icmp-ireq icmp-ireqreply icmp-maskreq icmp-maskreply |
Example 2-9 illustrates the use of Ethanalyzer to capture all packets hitting the inbound-low as well as inbound-hi queue on Nexus 6000. From the following outputs, notice that the TCP SYN/SYN ACK packets even for a BGP peering are part of the inbound-low queue, but the regular BGP updates and keepalives (such as the TCP packets after the BGP peering is established) and the acknowledgements are part of the inband-hi queue.
Example 2-9 Ethanalyzer Capture
N6k-1# ethanalyzer local interface inbound-low limit-captured-frames 20 Capturing on inband 2017-05-21 21:26:22.972623 10.162.223.33 -> 10.162.223.34 TCP bgp > 45912 [SYN, ACK] Seq=0 Ack=0 Win=16616 Len=0 MSS=1460 2017-05-21 21:26:33.214254 10.162.223.33 -> 10.162.223.34 TCP bgp > 14779 [SYN, ACK] Seq=0 Ack=0 Win=16616 Len=0 MSS=1460 2017-05-21 21:26:44.892236 8c:60:4f:a7:9a:6b -> 01:00:0c:cc:cc:cc CDP Device ID: N6k-1(FOC1934R1BF) Port ID: Ethernet1/4 2017-05-21 21:26:44.892337 8c:60:4f:a7:9a:68 -> 01:00:0c:cc:cc:cc CDP Device ID: N6k-1(FOC1934R1BF) Port ID: Ethernet1/1 2017-05-21 21:27:42.965431 00:25:45:e7:d0:00 -> 8c:60:4f:a7:9a:bc ARP 10.162.223 .34 is at 00:25:45:e7:d0:00 ! Output omitted for brevity N6k-1# ethanalyzer local interface inbound-hi limit-captured-frames 10 Capturing on inband 2017-05-21 21:34:42.821141 10.162.223.34 -> 10.162.223.33 BGP KEEPALIVE Message 2017-05-21 21:34:42.932217 10.162.223.33 -> 10.162.223.34 TCP bgp > 14779 [ACK] Seq=1 Ack=20 Win=17520 Len=0 2017-05-21 21:34:43.613048 10.162.223.33 -> 10.162.223.34 BGP KEEPALIVE Message 2017-05-21 21:34:43.814804 10.162.223.34 -> 10.162.223.33 TCP 14779 > bgp [ACK] Seq=20 Ack=20 Win=15339 Len=0 2017-05-21 21:34:46.005039 10.1.12.2 -> 224.0.0.5 OSPF Hello Packet 2017-05-21 21:34:46.919884 10.162.223.34 -> 10.162.223.33 BGP KEEPALIVE Message 2017-05-21 21:34:47.032215 10.162.223.33 -> 10.162.223.34 TCP bgp > 14779 [ACK] Seq=20 Ack=39 Win=17520 Len=0 ! Output omitted for brevity
As stated earlier, optimal practice is to write the captured frames in a file and then read it after the frames are captured. The saved file in a local bootflash is read using the command ethanalyzer local read location [detail].
Nexus 7000 offers no option for inbound-hi or inbound-low. The CLI supports captures on the mgmt interface or the inband interface. The inband interface captures both high- and low-priority packets. Example 2-10 illustrates how to write and read the saved packet capture data. In this example, Ethanalyzer is run with a capture-filter on STP packets.
Example 2-10 Ethanalyzer Write and Read
N7k-Admin# ethanalyzer local interface inband capture-filter "stp" write bootflash:stp.pcap Capturing on inband 10 N7k-Admin# ethanalyzer local read bootflash:stp.pcap 2017-05-21 23:48:30.216952 5c:fc:66:6c:f3:f6 -> Spanning-tree-(for-bridges)_00 STP 60 RST. Root = 4096/1/50:87:89:4b:bb:42 Cost = 0 Port = 0x9000 2017-05-21 23:48:30.426556 38:ed:18:a2:27:b0 -> Spanning-tree-(for-bridges)_00 STP 60 RST. Root = 4096/1/50:87:89:4b:bb:42 Cost = 1 Port = 0x8201 2017-05-21 23:48:30.426690 38:ed:18:a2:27:b0 -> Spanning-tree-(for-bridges)_00 STP 60 RST. Root = 4096/1/50:87:89:4b:bb:42 Cost = 1 Port = 0x8201 2017-05-21 23:48:30.426714 38:ed:18:a2:17:a6 -> Spanning-tree-(for-bridges)_00 ! Output omitted for brevity
! Detailed output of ethanalyzer N7k-Admin# ethanalyzer local read bootflash:stp.pcap detail Frame 1: 60 bytes on wire (480 bits), 60 bytes captured (480 bits) Encapsulation type: Ethernet (1) Arrival Time: May 21, 2017 23:48:30.216952000 UTC [Time shift for this packet: 0.000000000 seconds] Epoch Time: 1495410510.216952000 seconds [Time delta from previous captured frame: 0.000000000 seconds] [Time delta from previous displayed frame: 0.000000000 seconds] [Time since reference or first frame: 0.000000000 seconds] Frame Number: 1 Frame Length: 60 bytes (480 bits) Capture Length: 60 bytes (480 bits) [Frame is marked: False] [Frame is ignored: False] [Protocols in frame: eth:llc:stp] IEEE 802.3 Ethernet Destination: Spanning-tree-(for-bridges)_00 (01:80:c2:00:00:00) Address: Spanning-tree-(for-bridges)_00 (01:80:c2:00:00:00) .... ..0. .... .... .... .... = LG bit: Globally unique address (factory default) .... ...1 .... .... .... .... = IG bit: Group address (multicast/broadcast) Source: 5c:fc:66:6c:f3:f6 (5c:fc:66:6c:f3:f6) Address: 5c:fc:66:6c:f3:f6 (5c:fc:66:6c:f3:f6) .... ..0. .... .... .... .... = LG bit: Globally unique address (factory default) .... ...0 .... .... .... .... = IG bit: Individual address (unicast) Length: 39 Padding: 00000000000000 Logical-Link Control DSAP: Spanning Tree BPDU (0x42) IG Bit: Individual SSAP: Spanning Tree BPDU (0x42) CR Bit: Command Control field: U, func=UI (0x03) 000. 00.. = Command: Unnumbered Information (0x00) .... ..11 = Frame type: Unnumbered frame (0x03) Spanning Tree Protocol Protocol Identifier: Spanning Tree Protocol (0x0000) Protocol Version Identifier: Rapid Spanning Tree (2) BPDU Type: Rapid/Multiple Spanning Tree (0x02) BPDU flags: 0x3c (Forwarding, Learning, Port Role: Designated) 0... .... = Topology Change Acknowledgment: No .0.. .... = Agreement: No ..1. .... = Forwarding: Yes ...1 .... = Learning: Yes .... 11.. = Port Role: Designated (3) .... ..0. = Proposal: No .... ...0 = Topology Change: No Root Identifier: 4096 / 1 / 50:87:89:4b:bb:42 Root Bridge Priority: 4096 Root Bridge System ID Extension: 1 Root Bridge System ID: 50:87:89:4b:bb:42 (50:87:89:4b:bb:42) Root Path Cost: 0 Bridge Identifier: 4096 / 1 / 50:87:89:4b:bb:42 Bridge Priority: 4096 Bridge System ID Extension: 1 Bridge System ID: 50:87:89:4b:bb:42 (50:87:89:4b:bb:42) Port identifier: 0x9000 Message Age: 0 Max Age: 20 Hello Time: 2 Forward Delay: 15 Version 1 Length: 0 ! Output omitted for brevity
The saved .pcap file can also be transferred to a remote server via File Transfer Protocol (FTP), Trivial File Transfer Protocol (TFTP), Secure Copy Protocol (SCP), Secure FTP (SFTP), and Universal Serial Bus (USB), after which it can be easily analyzed using a packet analyzer tool such as Wireshark.
Packet Tracer
During troubleshooting, it becomes difficult to understand what action the system is taking on a particular packet or flow. For such instances, the packet tracer feature is used. Starting with NX-OS Version 7.0(3)I2(2a), the packet tracer utility was introduced on the Nexus 9000 switch. It is used when intermittent or complete packet loss is observed.
The packet tracer is configured in two simple steps:
Step 1. Define the filter.
Step 2. Start the packet tracer.
To set up the packet tracer, use the command test packet-tracer [src-ip src-ip | dst-ip dst-ip ] [protocol protocol-num | l4-src-port src-port | l4-dst-port dst-port]. Then start the packet tracer, using the command test packet-tracer start. To view the statistics of the specified traffic and the action on it, use the command test packet-tracer show. Finally, stop the packet tracer using the command test packet-tracer stop. Example 2-11 illustrates the use of the packet tracer to analyze the ICMP statistics between two hosts.
Example 2-11 Packet Tracer Configuration and Verification
! Defining the Filter in Packet-Tracer N9000-1# test packet-tracer src-ip 192.168.2.2 dst-ip 192.168.1.1 protocol 1 ! Starting the Packet-Tracer N9000-1# test packet-tracer start ! Verifying the statistics N9000-1# test packet-tracer show Packet-tracer stats --------------------- Module 1: Filter 1 installed: src-ip 192.168.2.2 dst-ip 192.168.1.1 protocol 1 ASIC instance 0: Entry 0: id = 9473, count = 120, active, fp, Entry 1: id = 9474, count = 0, active, hg, Filter 2 uninstalled: Filter 3 uninstalled: Filter 4 uninstalled: Filter 5 uninstalled: ! Second iteration of the Output N9000-1# test packet-tracer show Packet-tracer stats --------------------- Module 1: Filter 1 installed: src-ip 192.168.2.2 dst-ip 192.168.1.1 protocol 1 ASIC instance 0: Entry 0: id = 9473, count = 181, active, fp, Entry 1: id = 9474, count = 0, active, hg, Filter 2 uninstalled: Filter 3 uninstalled: Filter 4 uninstalled: Filter 5 uninstalled: ! Stopping the Packet-Tracer N9000-1# test packet-tracer stop
Even if the incoming traffic is dropped because of an ACL, the packet tracer helps determine whether the packet is reaching the router incoming interface. To remove all the filters from the packet tracer, use the command test packet-tracer remove-all.