Troubleshooting Guide

Overview

The Trellis Troubleshooting Tools (T3) is an ONOS application intended to help you troubleshoot traffic-related problems in your network by tracing installed flows and groups.

T3 works both on simulated OVS-based topologies but also on real HW, basing it’s internal decisions on the driver assigned to the device.

T3 also currently supports dual links between leaf and spine and properly handles the difference between deferred and immediate rules. Dual Homed hosts are also supported in T3.

To activate this tool on your ONOS environment, you need to run:

onos> app activate t3

T3 supports tracing of following packet types:

  • L2 Unicast

  • L2 Broadcast

  • L3 Unicast (both IPv4 and IPv6)

  • L3 Multicast (both IPv4 and IPv6)

  • LLDP and BDDP

  • ARP

  • DHCP

Network Information Base (NIB)

T3 analyzes network states that are stored in the NIB. A NIB is a collection of snapshots captured from various ONOS stores such as flows, groups, links…etc. The NIB, acting as a cache, allows T3 troubleshooting logic to avoid massive interaction with live ONOS subsystems, enabling near-runtime or postmortem analysis in a production environment.

There are two methods of loading the network snapshots into the NIB: (1) from a live running system or (2) from a onos-diagnostics file dump. Users can choose one load command before troubleshooting commands, or can run troubleshooting commands alone (then the NIB is automatically loaded via (1)).

t3-load-file

T3 offers this command to fill the NIB with files that represent snapshots of the network states in a certain time.

t3-load-file Help

onos> t3-load-file --help
DESCRIPTION
        onos:t3-load-file

        Load the NIB with onos-diagnostics dump files

SYNTAX
        onos:t3-load-file [options] rootDir

ARGUMENTS
        rootDir
                Specify the location of the directory where the previous onos-diagnostics command has extracted its dump files (e.g. /tmp/onos-diags/127.0.0.1)

OPTIONS
        --help
                Display this help message
        -j, --json
                Output JSON

t3-load-file Example

User is expected to run the tool onos-diagnostics 1 2 to generate target files first.

user$ onos-diagnostics -xj [IP_ONOS_INSTANCE_1 IP_ONOS_INSTANCE_2 ...]
Collecting diagnostics on [IP_ONOS_INSTANCE_1]...logs netcfg.json devices.json device-drivers.json links.json hosts.json ports.json edge-ports.json flows.json groups.json masters.json routes.json mcast-host-show.json  Done.
Collecting diagnostics on [IP_ONOS_INSTANCE_2]...logs netcfg.json devices.json device-drivers.json links.json hosts.json ports.json edge-ports.json flows.json groups.json masters.json routes.json mcast-host-show.json  Done.

Then, execute this command with the path of the location where those JSON files are extracted. User can check the creation time and a summary of the NIB contents.

onos> t3-load-file /tmp/onos-diags/[IP_ONOS_INSTANCE_TARGET]
Load target files in: /tmp/onos-diags/[IP_ONOS_INSTANCE_TARGET]/
FlowNib created 03-06-2020 01:30 from FILE
GroupNib created 03-06-2020 01:30 from FILE
LinkNib created 03-06-2020 01:30 from FILE
HostNib created 03-06-2020 01:30 from FILE
DeviceNib created 03-06-2020 01:30 from FILE
DriverNib created 03-06-2020 01:30 from FILE
MastershipNib created 03-06-2020 01:30 from FILE
EdgePortNib created 03-06-2020 01:30 from FILE
RouteNib created 03-06-2020 01:30 from FILE
NetworkConfigNib created 03-06-2020 01:30 from FILE
MulticastRouteNib created 03-06-2020 01:30 from FILE

Alternatively, it is possible to use onos-diagnostics-k8s 3 2 in Kubernetes enabled environments.

t3-load-snapshot

T3 offers this command to fill the NIB with snapshots of the network states at the execution time.

t3-load-snapshot Help

onos> t3-load-snapshot --help
DESCRIPTION
        onos:t3-load-snapshot

        Load the NIB with the network states stored in the ONOS instance where the T3 is running

SYNTAX
        onos:t3-load-snapshot [options]

OPTIONS
        --help
                Display this help message
        -j, --json
                Output JSON

t3-load-snapshot Example

User can check the creation time and a summary of the NIB contents. SNAPSHOT indicates that the source of the current NIB is the ONOS instance where the T3 is running.

onos> t3-load-snapshot
Load current network states from ONOS stores
FlowNib created 03-06-2020 01:40 from SNAPSHOT
GroupNib created 03-06-2020 01:40 from SNAPSHOT
LinkNib created 03-06-2020 01:40 from SNAPSHOT
HostNib created 03-06-2020 01:40 from SNAPSHOT
DeviceNib created 03-06-2020 01:40 from SNAPSHOT
DriverNib created 03-06-2020 01:40 from SNAPSHOT
MastershipNib created 03-06-2020 01:40 from SNAPSHOT
EdgePortNib created 03-06-2020 01:40 from SNAPSHOT
RouteNib created 03-06-2020 01:40 from SNAPSHOT
NetworkConfigNib created 03-06-2020 01:40 from SNAPSHOT
MulticastRouteNib created 03-06-2020 01:40 from SNAPSHOT

If one of T3 commands below is used before loading the NIB from SNAPSHOT or FILE (so the NIB is empty), the command internally executes t3-load-snapshot to proceed with auto-filled NIB.

onos> t3-troubleshoot-simple 00:AA:00:00:00:01/None 00:AA:00:00:00:02/None
*** NIB is invalid. Snapshots for the NIB have been auto-filled: ***
Load current network states from ONOS stores
FlowNib created 03-06-2020 01:44 from SNAPSHOT
GroupNib created 03-06-2020 01:44 from SNAPSHOT
LinkNib created 03-06-2020 01:44 from SNAPSHOT
...
Tracing between: 00:AA:00:00:00:01/None and 00:AA:00:00:00:02/None
Tracing Packet: DefaultTrafficSelector{criteria=[IN_PORT:3, ETH_DST:00:00:00:00:02:04, ETH_TYPE:ipv4, VLAN_VID:None, IPV4_SRC:10.0.2.1/32, IPV4_DST:10.0.2.2/32]}
Paths
[of:0000000000000204/3, of:0000000000000204/4]
Result:
Reached required destination Host of:0000000000000204/4

t3-troubleshoot

As of right now the tools provides a CLI command (troubleshoot).

As arguments to the CLI command you have to pass arguments describing the packet you want to trace through the network.

t3-troubleshoot Help

onos> t3-troubleshoot --help
DESCRIPTION
        onos:troubleshoot

    troubleshoots flows and groups between source and destination

SYNTAX
        onos:troubleshoot [options]

OPTIONS
        -sm, --srcMac
                Source MAC
        -j, --json
                Output JSON
        --help
                Display this help message
        -v, --verbose
                Outputs complete path
        -ipp, --ipProto
                IP Proto
        -s, --srcIp
                Source IP
        -dm, --dstMac
                Destination MAC
        -ml, --mplsLabel
                Mpls label of incoming packet
        -vv, --veryverbose
                Outputs flows and groups for every device
        -udps, --udpSrc
                UDP Source
        -mb, --mplsBos
                MPLS BOS
        -et, --ethType
                ETH Type
                (defaults to ipv4)
        -sp, --srcPort
                Source Port
        -dtp, --dstTcpPort
                destination TCP Port
        -udpd, --udpDst
                UDP Destination
        -stp, --srcTcpPort
                Source TCP Port
        -d, --dstIp
                Destination IP
        -vid, --vlanId
                Vlan of incoming packet
                (defaults to None)

Default Verbosity Level

Let’s look at an example.

onos> t3-troubleshoot -s 10.0.2.1 -sp of:0000000000000204/3 -sm 00:AA:00:00:00:01 -et IPV4 -d 10.0.3.1 -dm 00:00:00:00:02:04 -vid None

with this command we are asking the t3 tool to trace a packet with

  • Source IP of 10.0.2.1/32

  • Input port of:0000000000000204/3

  • Source MAC 00:AA:00:00:00:01

  • ethType IPv4

  • Source TCP Port 12345

  • Destination IP 10.0.3.1/32

  • Destination MAC 00:AA:00:00:00:01

  • Destination TCP Port 80 and

  • VLAN ID None

This command has a verbosity level of 0 and will output just the paths that the packet might take inside the network. In the example deployment taken from trellis.py 4, the output will be:

Tracing packet: [IN_PORT:3, ETH_DST:00:00:00:00:02:04, ETH_SRC:00:AA:00:00:00:01, ETH_TYPE:ipv4, VLAN_VID:None, IPV4_SRC:10.0.2.1/32, IPV4_DST:10.0.3.1/32, TCP_SRC:12345, TCP_DST:80, MPLS_BOS:true]
Paths
[of:0000000000000204/3, of:0000000000000204/2, of:0000000000000227/1, of:0000000000000227/2, of:0000000000000205/2, of:0000000000000205/3]
[of:0000000000000204/3, of:0000000000000204/1, of:0000000000000226/1, of:0000000000000226/2, of:0000000000000205/1, of:0000000000000205/3]
Result:
Reached required destination Host
Reached required destination Host

Middle Verbosity Level

There are also two other verbosity levels.

onos> t3-troubleshoot -v -s 10.0.2.1 -sp of:0000000000000204/3 -sm 00:AA:00:00:00:01 -et IPV4 -d 10.0.3.1 -dm 00:00:00:00:02:04 -vid None

prints flows selectors and groupId that are met along the path

Tracing packet: [IN_PORT:3, ETH_DST:00:00:00:00:02:04, ETH_SRC:00:AA:00:00:00:01, ETH_TYPE:ipv4, VLAN_VID:None, IPV4_SRC:10.0.2.1/32, IPV4_DST:10.0.3.1/32, TCP_SRC:12345, TCP_DST:80, MPLS_BOS:true]
Path [of:0000000000000204/3, of:0000000000000204/2, of:0000000000000227/1, of:0000000000000227/2, of:0000000000000205/2, of:0000000000000205/3]
Device of:0000000000000204
Input from of:0000000000000204/3
Flows
flowId=a20000bd57cb49, selector=[]
flowId=900000d4173ff7, selector=[IN_PORT:3, VLAN_VID:None]
flowId=9000004b3d1965, selector=[IN_PORT:3, ETH_DST:00:00:00:00:02:04, ETH_TYPE:ipv4, VLAN_VID:20]
flowId=900000499769f5, selector=[ETH_TYPE:ipv4, IPV4_DST:10.0.3.0/24]
flowId=a2000038b487e7, selector=[]
Groups
groupId=GroupId{id=0x70000010}
groupId=GroupId{id=0x9200000d}
groupId=GroupId{id=0x9000000c}
groupId=GroupId{id=0xffe0002}
groupId=GroupId{id=0x9200000f}
groupId=GroupId{id=0x9000000e}
groupId=GroupId{id=0xffe0001}
Output through of:0000000000000204/2
Device of:0000000000000227
Input from of:0000000000000227/1
Flows
flowId=a200005577bcd6, selector=[]
flowId=900000498457b2, selector=[IN_PORT:1, VLAN_VID:None]
flowId=900000f87746c9, selector=[IN_PORT:1, ETH_DST:00:00:00:00:02:27, ETH_TYPE:mpls_unicast, VLAN_VID:4094]
flowId=a200005bfd6ed7, selector=[]
flowId=9000009f3d554f, selector=[ETH_TYPE:mpls_unicast, MPLS_LABEL:205, MPLS_BOS:true]
flowId=a20000e601fc35, selector=[]
Groups
groupId=GroupId{id=0x70000017}
groupId=GroupId{id=0x20000016}
groupId=GroupId{id=0xffe0002}
Output through of:0000000000000227/2
Device of:0000000000000205
Input from of:0000000000000205/2
Flows
flowId=a200008d542817, selector=[]
flowId=900000be3179a0, selector=[IN_PORT:1, VLAN_VID:None]
flowId=900000766be164, selector=[IN_PORT:1, ETH_DST:00:00:00:00:02:05, ETH_TYPE:ipv4, VLAN_VID:4094]
flowId=900000ec1d5779, selector=[ETH_TYPE:ipv4, IPV4_DST:10.0.3.1/32]
flowId=a2000042d8e6b5, selector=[]
Groups
groupId=GroupId{id=0x2000002a}
groupId=GroupId{id=0x1e0003}
groupId=GroupId{id=0x2000002a}
groupId=GroupId{id=0x1e0003}
Output through of:0000000000000205/3
Path [of:0000000000000204/3, of:0000000000000204/1, of:0000000000000226/1, of:0000000000000226/2, of:0000000000000205/1, of:0000000000000205/3]
Device of:0000000000000204
Input from of:0000000000000204/3
Flows
flowId=a20000bd57cb49, selector=[]
flowId=900000d4173ff7, selector=[IN_PORT:3, VLAN_VID:None]
flowId=9000004b3d1965, selector=[IN_PORT:3, ETH_DST:00:00:00:00:02:04, ETH_TYPE:ipv4, VLAN_VID:20]
flowId=900000499769f5, selector=[ETH_TYPE:ipv4, IPV4_DST:10.0.3.0/24]
flowId=a2000038b487e7, selector=[]
Groups
groupId=GroupId{id=0x70000010}
groupId=GroupId{id=0x9200000d}
groupId=GroupId{id=0x9000000c}
groupId=GroupId{id=0xffe0002}
groupId=GroupId{id=0x9200000f}
groupId=GroupId{id=0x9000000e}
groupId=GroupId{id=0xffe0001}
Output through of:0000000000000204/1
Device of:0000000000000226
Input from of:0000000000000226/1
Flows
flowId=a20000e9b27d24, selector=[]
flowId=9000004bd76105, selector=[IN_PORT:1, VLAN_VID:None]
flowId=90000069f274d1, selector=[IN_PORT:1, ETH_DST:00:00:00:00:02:26, ETH_TYPE:mpls_unicast, VLAN_VID:4094]
flowId=a200000ec7529e, selector=[]
flowId=90000050f9a350, selector=[ETH_TYPE:mpls_unicast, MPLS_LABEL:205, MPLS_BOS:true]
flowId=a20000b15cada6, selector=[]
Groups
groupId=GroupId{id=0x7000000b}
groupId=GroupId{id=0x20000009}
groupId=GroupId{id=0xffe0002}
Output through of:0000000000000226/2
Device of:0000000000000205
Input from of:0000000000000205/1
Flows
flowId=a200008d542817, selector=[]
flowId=900000be3179a0, selector=[IN_PORT:1, VLAN_VID:None]
flowId=900000766be164, selector=[IN_PORT:1, ETH_DST:00:00:00:00:02:05, ETH_TYPE:ipv4, VLAN_VID:4094]
flowId=900000ec1d5779, selector=[ETH_TYPE:ipv4, IPV4_DST:10.0.3.1/32]
flowId=a2000042d8e6b5, selector=[]
Groups
groupId=GroupId{id=0x2000002a}
groupId=GroupId{id=0x1e0003}
groupId=GroupId{id=0x2000002a}
groupId=GroupId{id=0x1e0003}
Output through of:0000000000000205/3
Result:
Reached required destination Host
Reached required destination Host

High Verbosity Level

The most verbose output you can get is

onos> t3-troubleshoot -vv -s 10.0.2.1 -sp of:0000000000000204/3 -sm 00:AA:00:00:00:01 -et IPV4 -d 10.0.3.1 -dm 00:00:00:00:02:04 -vid None

This verbosity level prints all the flows and groups that are met along the path:

Tracing packet: [IN_PORT:3, ETH_DST:00:00:00:00:02:04, ETH_SRC:00:AA:00:00:00:01, ETH_TYPE:ipv4, VLAN_VID:None, IPV4_SRC:10.0.2.1/32, IPV4_DST:10.0.3.1/32, TCP_SRC:12345, TCP_DST:80, MPLS_BOS:true]
Path [of:0000000000000204/3, of:0000000000000204/2, of:0000000000000227/1, of:0000000000000227/2, of:0000000000000205/2, of:0000000000000205/3]
Device of:0000000000000204
Input from of:0000000000000204/3
Flows
ADDED, bytes=733506, packets=8691, table=0, priority=0, selector=[], treatment=[transition=TABLE:10]
ADDED, bytes=11110, packets=61, table=10, priority=32768, selector=[IN_PORT:3, VLAN_VID:None], treatment=[immediate=[VLAN_PUSH:vlan, VLAN_ID:20], transition=TABLE:20]
ADDED, bytes=8698, packets=29, table=20, priority=32768, selector=[IN_PORT:3, ETH_DST:00:00:00:00:02:04, ETH_TYPE:ipv4, VLAN_VID:20], treatment=[transition=TABLE:30]
ADDED, bytes=16612, packets=50, table=30, priority=48010, selector=[ETH_TYPE:ipv4, IPV4_DST:10.0.3.0/24], treatment=[deferred=[GROUP:0x70000010], transition=TABLE:60]
ADDED, bytes=34400, packets=112, table=60, priority=0, selector=[], treatment=[immediate=[NOACTION]]
Groups
id=0x70000010, state=ADDED, type=SELECT, bytes=16906, packets=53, appId=org.onosproject.segmentrouting, referenceCount=3
id=0x70000010, bucket=1, bytes=490, packets=5, actions=[GROUP:0x9200000d]
id=0x70000010, bucket=2, bytes=16416, packets=48, actions=[GROUP:0x9200000f]
id=0x9200000d, state=ADDED, type=INDIRECT, bytes=490, packets=5, appId=org.onosproject.segmentrouting, referenceCount=0
id=0x9200000d, bucket=1, bytes=490, packets=5, actions=[VLAN_POP, MPLS_PUSH:mpls_unicast, MPLS_LABEL:205, GROUP:0x9000000c, VLAN_PUSH:vlan, VLAN_ID:4095]
id=0x9000000c, state=ADDED, type=INDIRECT, bytes=490, packets=5, appId=org.onosproject.segmentrouting, referenceCount=0
id=0x9000000c, bucket=1, bytes=490, packets=5, actions=[ETH_DST:00:00:00:00:02:27, ETH_SRC:00:00:00:00:02:04, VLAN_ID:4094, GROUP:0xffe0002]
id=0xffe0002, state=ADDED, type=INDIRECT, bytes=726, packets=7, appId=org.onosproject.segmentrouting, referenceCount=0
id=0xffe0002, bucket=1, bytes=726, packets=7, actions=[VLAN_POP, OUTPUT:2]
id=0x9200000f, state=ADDED, type=INDIRECT, bytes=16416, packets=48, appId=org.onosproject.segmentrouting, referenceCount=0
id=0x9200000f, bucket=1, bytes=16416, packets=48, actions=[VLAN_POP, MPLS_PUSH:mpls_unicast, MPLS_LABEL:205, GROUP:0x9000000e, VLAN_PUSH:vlan, VLAN_ID:4095]
id=0x9000000e, state=ADDED, type=INDIRECT, bytes=16416, packets=48, appId=org.onosproject.segmentrouting, referenceCount=0
id=0x9000000e, bucket=1, bytes=16416, packets=48, actions=[ETH_DST:00:00:00:00:02:26, ETH_SRC:00:00:00:00:02:04, VLAN_ID:4094, GROUP:0xffe0001]
id=0xffe0001, state=ADDED, type=INDIRECT, bytes=16416, packets=48, appId=org.onosproject.segmentrouting, referenceCount=0
id=0xffe0001, bucket=1, bytes=16416, packets=48, actions=[VLAN_POP, OUTPUT:1]
Output through of:0000000000000204/2
Device of:0000000000000227
Input from of:0000000000000227/1
Flows
ADDED, bytes=702318, packets=8511, table=0, priority=0, selector=[], treatment=[transition=TABLE:10]
ADDED, bytes=343130, packets=4234, table=10, priority=32768, selector=[IN_PORT:1, VLAN_VID:None], treatment=[immediate=[VLAN_PUSH:vlan, VLAN_ID:4094], transition=TABLE:20]
ADDED, bytes=754, packets=7, table=20, priority=32768, selector=[IN_PORT:1, ETH_DST:00:00:00:00:02:27, ETH_TYPE:mpls_unicast, VLAN_VID:4094], treatment=[transition=TABLE:23]
ADDED, bytes=17566, packets=57, table=23, priority=0, selector=[], treatment=[transition=TABLE:24]
ADDED, bytes=510, packets=5, table=24, priority=100, selector=[ETH_TYPE:mpls_unicast, MPLS_LABEL:205, MPLS_BOS:true], treatment=[deferred=[MPLS_POP:ipv4, DEC_TTL, GROUP:0x70000017], transition=TABLE:60]
ADDED, bytes=17566, packets=57, table=60, priority=0, selector=[], treatment=[immediate=[NOACTION]]
Groups
id=0x70000017, state=ADDED, type=SELECT, bytes=754, packets=7, appId=org.onosproject.segmentrouting, referenceCount=2
id=0x70000017, bucket=1, bytes=754, packets=7, actions=[GROUP:0x20000016]
id=0x20000016, state=ADDED, type=INDIRECT, bytes=754, packets=7, appId=org.onosproject.segmentrouting, referenceCount=0
id=0x20000016, bucket=1, bytes=754, packets=7, actions=[ETH_DST:00:00:00:00:02:05, ETH_SRC:00:00:00:00:02:27, VLAN_ID:4094, GROUP:0xffe0002]
id=0xffe0002, state=ADDED, type=INDIRECT, bytes=754, packets=7, appId=org.onosproject.segmentrouting, referenceCount=0
id=0xffe0002, bucket=1, bytes=754, packets=7, actions=[VLAN_POP, OUTPUT:2]
Output through of:0000000000000227/2
Device of:0000000000000205
Input from of:0000000000000205/2
Flows
ADDED, bytes=2105886, packets=24084, table=0, priority=0, selector=[], treatment=[transition=TABLE:10]
ADDED, bytes=358792, packets=4275, table=10, priority=32768, selector=[IN_PORT:1, VLAN_VID:None], treatment=[immediate=[VLAN_PUSH:vlan, VLAN_ID:4094], transition=TABLE:20]
ADDED, bytes=16416, packets=48, table=20, priority=32768, selector=[IN_PORT:1, ETH_DST:00:00:00:00:02:05, ETH_TYPE:ipv4, VLAN_VID:4094], treatment=[transition=TABLE:30]
ADDED, bytes=196, packets=2, table=30, priority=64010, selector=[ETH_TYPE:ipv4, IPV4_DST:10.0.3.1/32], treatment=[deferred=[GROUP:0x2000002a], transition=TABLE:60]
ADDED, bytes=1370884, packets=15163, table=60, priority=0, selector=[], treatment=[immediate=[NOACTION]]
Groups
id=0x2000002a, state=ADDED, type=INDIRECT, bytes=196, packets=2, appId=org.onosproject.segmentrouting, referenceCount=1
id=0x2000002a, bucket=1, bytes=196, packets=2, actions=[ETH_DST:00:AA:00:00:00:03, ETH_SRC:00:00:00:00:02:05, VLAN_ID:30, GROUP:0x1e0003]
id=0x1e0003, state=ADDED, type=INDIRECT, bytes=12960, packets=126, appId=org.onosproject.segmentrouting, referenceCount=1
id=0x1e0003, bucket=1, bytes=12960, packets=126, actions=[VLAN_POP, OUTPUT:3]
id=0x2000002a, state=ADDED, type=INDIRECT, bytes=196, packets=2, appId=org.onosproject.segmentrouting, referenceCount=1
id=0x2000002a, bucket=1, bytes=196, packets=2, actions=[ETH_DST:00:AA:00:00:00:03, ETH_SRC:00:00:00:00:02:05, VLAN_ID:30, GROUP:0x1e0003]
id=0x1e0003, state=ADDED, type=INDIRECT, bytes=12960, packets=126, appId=org.onosproject.segmentrouting, referenceCount=1
id=0x1e0003, bucket=1, bytes=12960, packets=126, actions=[VLAN_POP, OUTPUT:3]
Output through of:0000000000000205/3
Path [of:0000000000000204/3, of:0000000000000204/1, of:0000000000000226/1, of:0000000000000226/2, of:0000000000000205/1, of:0000000000000205/3]
Device of:0000000000000204
Input from of:0000000000000204/3
Flows
ADDED, bytes=733506, packets=8691, table=0, priority=0, selector=[], treatment=[transition=TABLE:10]
ADDED, bytes=11110, packets=61, table=10, priority=32768, selector=[IN_PORT:3, VLAN_VID:None], treatment=[immediate=[VLAN_PUSH:vlan, VLAN_ID:20], transition=TABLE:20]
ADDED, bytes=8698, packets=29, table=20, priority=32768, selector=[IN_PORT:3, ETH_DST:00:00:00:00:02:04, ETH_TYPE:ipv4, VLAN_VID:20], treatment=[transition=TABLE:30]
ADDED, bytes=16612, packets=50, table=30, priority=48010, selector=[ETH_TYPE:ipv4, IPV4_DST:10.0.3.0/24], treatment=[deferred=[GROUP:0x70000010], transition=TABLE:60]
ADDED, bytes=34400, packets=112, table=60, priority=0, selector=[], treatment=[immediate=[NOACTION]]
Groups
id=0x70000010, state=ADDED, type=SELECT, bytes=16906, packets=53, appId=org.onosproject.segmentrouting, referenceCount=3
id=0x70000010, bucket=1, bytes=490, packets=5, actions=[GROUP:0x9200000d]
id=0x70000010, bucket=2, bytes=16416, packets=48, actions=[GROUP:0x9200000f]
id=0x9200000d, state=ADDED, type=INDIRECT, bytes=490, packets=5, appId=org.onosproject.segmentrouting, referenceCount=0
id=0x9200000d, bucket=1, bytes=490, packets=5, actions=[VLAN_POP, MPLS_PUSH:mpls_unicast, MPLS_LABEL:205, GROUP:0x9000000c, VLAN_PUSH:vlan, VLAN_ID:4095]
id=0x9000000c, state=ADDED, type=INDIRECT, bytes=490, packets=5, appId=org.onosproject.segmentrouting, referenceCount=0
id=0x9000000c, bucket=1, bytes=490, packets=5, actions=[ETH_DST:00:00:00:00:02:27, ETH_SRC:00:00:00:00:02:04, VLAN_ID:4094, GROUP:0xffe0002]
id=0xffe0002, state=ADDED, type=INDIRECT, bytes=726, packets=7, appId=org.onosproject.segmentrouting, referenceCount=0
id=0xffe0002, bucket=1, bytes=726, packets=7, actions=[VLAN_POP, OUTPUT:2]
id=0x9200000f, state=ADDED, type=INDIRECT, bytes=16416, packets=48, appId=org.onosproject.segmentrouting, referenceCount=0
id=0x9200000f, bucket=1, bytes=16416, packets=48, actions=[VLAN_POP, MPLS_PUSH:mpls_unicast, MPLS_LABEL:205, GROUP:0x9000000e, VLAN_PUSH:vlan, VLAN_ID:4095]
id=0x9000000e, state=ADDED, type=INDIRECT, bytes=16416, packets=48, appId=org.onosproject.segmentrouting, referenceCount=0
id=0x9000000e, bucket=1, bytes=16416, packets=48, actions=[ETH_DST:00:00:00:00:02:26, ETH_SRC:00:00:00:00:02:04, VLAN_ID:4094, GROUP:0xffe0001]
id=0xffe0001, state=ADDED, type=INDIRECT, bytes=16416, packets=48, appId=org.onosproject.segmentrouting, referenceCount=0
id=0xffe0001, bucket=1, bytes=16416, packets=48, actions=[VLAN_POP, OUTPUT:1]
Output through of:0000000000000204/1
Device of:0000000000000226
Input from of:0000000000000226/1
Flows
ADDED, bytes=701910, packets=8507, table=0, priority=0, selector=[], treatment=[transition=TABLE:10]
ADDED, bytes=358984, packets=4275, table=10, priority=32768, selector=[IN_PORT:1, VLAN_VID:None], treatment=[immediate=[VLAN_PUSH:vlan, VLAN_ID:4094], transition=TABLE:20]
ADDED, bytes=16608, packets=48, table=20, priority=32768, selector=[IN_PORT:1, ETH_DST:00:00:00:00:02:26, ETH_TYPE:mpls_unicast, VLAN_VID:4094], treatment=[transition=TABLE:23]
ADDED, bytes=17158, packets=53, table=23, priority=0, selector=[], treatment=[transition=TABLE:24]
ADDED, bytes=16608, packets=48, table=24, priority=100, selector=[ETH_TYPE:mpls_unicast, MPLS_LABEL:205, MPLS_BOS:true], treatment=[deferred=[MPLS_POP:ipv4, DEC_TTL, GROUP:0x7000000b], transition=TABLE:60]
ADDED, bytes=17158, packets=53, table=60, priority=0, selector=[], treatment=[immediate=[NOACTION]]
Groups
id=0x7000000b, state=ADDED, type=SELECT, bytes=16608, packets=48, appId=org.onosproject.segmentrouting, referenceCount=2
id=0x7000000b, bucket=1, bytes=16608, packets=48, actions=[GROUP:0x20000009]
id=0x20000009, state=ADDED, type=INDIRECT, bytes=16608, packets=48, appId=org.onosproject.segmentrouting, referenceCount=0
id=0x20000009, bucket=1, bytes=16608, packets=48, actions=[ETH_DST:00:00:00:00:02:05, ETH_SRC:00:00:00:00:02:26, VLAN_ID:4094, GROUP:0xffe0002]
id=0xffe0002, state=ADDED, type=INDIRECT, bytes=16608, packets=48, appId=org.onosproject.segmentrouting, referenceCount=0
id=0xffe0002, bucket=1, bytes=16608, packets=48, actions=[VLAN_POP, OUTPUT:2]
Output through of:0000000000000226/2
Device of:0000000000000205
Input from of:0000000000000205/1
Flows
ADDED, bytes=2105886, packets=24084, table=0, priority=0, selector=[], treatment=[transition=TABLE:10]
ADDED, bytes=358792, packets=4275, table=10, priority=32768, selector=[IN_PORT:1, VLAN_VID:None], treatment=[immediate=[VLAN_PUSH:vlan, VLAN_ID:4094], transition=TABLE:20]
ADDED, bytes=16416, packets=48, table=20, priority=32768, selector=[IN_PORT:1, ETH_DST:00:00:00:00:02:05, ETH_TYPE:ipv4, VLAN_VID:4094], treatment=[transition=TABLE:30]
ADDED, bytes=196, packets=2, table=30, priority=64010, selector=[ETH_TYPE:ipv4, IPV4_DST:10.0.3.1/32], treatment=[deferred=[GROUP:0x2000002a], transition=TABLE:60]
ADDED, bytes=1370884, packets=15163, table=60, priority=0, selector=[], treatment=[immediate=[NOACTION]]
Groups
id=0x2000002a, state=ADDED, type=INDIRECT, bytes=196, packets=2, appId=org.onosproject.segmentrouting, referenceCount=1
id=0x2000002a, bucket=1, bytes=196, packets=2, actions=[ETH_DST:00:AA:00:00:00:03, ETH_SRC:00:00:00:00:02:05, VLAN_ID:30, GROUP:0x1e0003]
id=0x1e0003, state=ADDED, type=INDIRECT, bytes=12960, packets=126, appId=org.onosproject.segmentrouting, referenceCount=1
id=0x1e0003, bucket=1, bytes=12960, packets=126, actions=[VLAN_POP, OUTPUT:3]
id=0x2000002a, state=ADDED, type=INDIRECT, bytes=196, packets=2, appId=org.onosproject.segmentrouting, referenceCount=1
id=0x2000002a, bucket=1, bytes=196, packets=2, actions=[ETH_DST:00:AA:00:00:00:03, ETH_SRC:00:00:00:00:02:05, VLAN_ID:30, GROUP:0x1e0003]
id=0x1e0003, state=ADDED, type=INDIRECT, bytes=12960, packets=126, appId=org.onosproject.segmentrouting, referenceCount=1
id=0x1e0003, bucket=1, bytes=12960, packets=126, actions=[VLAN_POP, OUTPUT:3]
Output through of:0000000000000205/3
Result:
Reached required destination Host
Reached required destination Host

Successful Example

In this example we trace a packet that should be bridged between two hosts in the same subnet

onos> t3-troubleshoot -s 10.0.2.1 -sp of:0000000000000204/3 -sm 00:AA:00:00:00:01 -et IPV4 -d 10.0.3.1 -dm 00:AA:00:00:00:02 -vid None
Tracing packet: [IN_PORT:3, ETH_DST:00:AA:00:00:00:02, ETH_SRC:00:AA:00:00:00:01, ETH_TYPE:ipv4, VLAN_VID:None, IPV4_SRC:10.0.2.1/32, IPV4_DST:10.0.2.2/32, TCP_SRC:12345, TCP_DST:80, MPLS_BOS:true]
Paths
[of:0000000000000204/3, of:0000000000000204/4]
Result:
Reached required destination Host

Failed Example

Obviously failures can happen along one of the paths that the packet can take.

In our sample we removed a flow in of:0000000000000226 thus disabling that possible path for our initial packet.

The output of our trace would then look something like:

onos> t3-troubleshoot -s 10.0.2.1 -sp of:0000000000000204/3 -sm 00:AA:00:00:00:01 -et IPV4 -d 10.0.3.1 -dm 00:00:00:00:02:04 -vid None
Tracing packet: [IN_PORT:3, ETH_DST:00:00:00:00:02:04, ETH_SRC:00:AA:00:00:00:01, ETH_TYPE:ipv4, VLAN_VID:None, IPV4_SRC:10.0.2.1/32, IPV4_DST:10.0.3.1/32, TCP_SRC:12345, TCP_DST:80, MPLS_BOS:true]
Paths
[of:0000000000000204/3, of:0000000000000204/2, of:0000000000000227/1, of:0000000000000227/2, of:0000000000000205/2, of:0000000000000205/3]
[of:0000000000000204/3, of:0000000000000204/1, of:0000000000000226/1]
Result:
Reached required destination Host
No output out of device of:0000000000000226 Packet is dropped

Notice that on the second path that was previously going through of:0000000000000226 now the packet is dropped due to the rule we removed.

Example of Various Packet Types

These sample commands refer to the Mininet environment found in trellis.py 4

  • L2 Unicast IPv4

    t3-troubleshoot -s 10.0.2.1 -sp of:0000000000000204/3 -sm 00:AA:00:00:00:01 -d 10.0.2.2 -dm 00:AA:00:00:00:02

  • L2 Unicast IPv6

    t3-troubleshoot -vv -s 2000::301  -sp of:0000000000000205/5 -sm 00:BB:00:00:00:03 -d 2000::302 -dm 00:BB:00:00:00:04 -et ipv6

  • L3 Unicast IPv4

    t3-troubleshoot -s 10.0.2.1 -sp of:0000000000000204/3 -sm 00:AA:00:00:00:01 -d 10.0.3.1 -dm 00:00:00:00:02:04

  • L3 Unicast IPv6

    t3-troubleshoot -vv -s 2000::301  -sp of:0000000000000205/5 -sm 00:BB:00:00:00:03 -d 2000::202 -dm 00:00:00:00:02:05 -et ipv6

  • ARP

    t3-troubleshoot -sp of:0000000000000204/3 -et ARP

  • DHCP

    t3-troubleshoot -vv -s 10.0.2.1 -sp of:0000000000000204/3 -sm 00:AA:00:00:00:01 -d 10.0.3.253 -dm 00:00:00:00:02:04 -ipp 17 -udps 67 -udpd 67

  • LLDP

    t3-troubleshoot -sp of:0000000000000204/3 -et lldp

  • BDDP

    t3-troubleshoot -sp of:0000000000000204/3 -et bddp

  • LLDP from CONTROLLER

    T3 is also capable of tracing packets coming in from the CONTROLLER.

    We use the same input port with the possibility to give CONTROLLER as logical port number.

    This is a mere convenience to carry the information inside T3 and do the proper flooding to all the ports of the given device.

    This case is not treated like all the rest and the input port here is a logical entity that has no match on any physical element on the device.

    t3-troubleshoot -sp of:0000000000000204/CONTROLLER -et lldp

  • Multicast

    T3 is also capable of tracing multicast packets.

    t3-troubleshoot -vv -sp of:0000000000000205/9 -et ipv4 -d 224.0.0.1 -dm 01:00:5e:00:00:01

t3-troubleshoot-simple

T3 also offers a simplified version of the command to be able to ping between two given hosts without having to specify all the information.

This commands bases itself on top of some configuration present in ONOS through netcfg, namely the VLAN interface configuration and the segment routing leaf router MAC for L3 Unicast. If that is not present the user needs to input this information manually recurring to the the previously shown command. Same verbosity levels and output as the troubleshoot command are possible.

t3-troubleshoot-simple Help

onos> t3-troubleshoot-simple --help
DESCRIPTION
        onos:troubleshoot-simple

    Given two hosts troubleshoots flows and groups between them, in case of segment routing

SYNTAX
        onos:troubleshoot-simple [options] one two

ARGUMENTS
        one
                One host ID
        two
                Another host ID

OPTIONS
        -vv, --veryverbose
                Outputs flows and groups for every device
        -j, --json
                Output JSON
        --help
                Display this help message
        -et, --ethType
                ETH Type
                (defaults to ipv4)
        -v, --verbose
                Outputs complete path

t3-troubleshoot-simple Example

Examples are given for the same trellis.py 4 script.

  • L2 Unicast

    Hosts are under the same leaf, only interface configuration is needed.

    t3-troubleshoot-simple -v -et IPV4 00:AA:00:00:00:01/None 00:AA:00:00:00:02/None

  • L3 Multicast

    Hosts are under different leafs, also segment routing configuration config is needed for the leaf’s routerMac.

    t3-troubleshoot-simple -et IPV4 00:AA:00:00:00:01/None 00:AA:00:00:00:03/None

t3-troubleshoot-pingall

T3 also offers a ping-all command to be able to test ping between any given hosts of an EthType. This commands bases itself on top of the simplified Host to Host command. It has two verbosity levels, one just with Host and Destination with the result message of the Trace and one that shows also the packet and the paths.

t3-troubleshoot-pingall Help

onos> t3-troubleshoot-pingall --help
DESCRIPTION
        onos:t3-troubleshoot-pingall

    Traces a ping between all hosts in the system of a given ETH type

SYNTAX
        onos:t3-troubleshoot-pingall [options]

OPTIONS
        -j, --json
                Output JSON
        --help
                Display this help message
        -et, --ethType
                ETH Type
                (defaults to ipv4)
        -v, --verbose
                Outputs trace for each host to host combination

t3-troubleshoot-pingall Examples

  • IPv4 hosts

    t3-troubleshoot-pingall

  • IPv6 hosts

    t3-troubleshoot-pingall -et IPV6

t3-troubleshoot-mcast

T3 also offers a mcast command queries the system for all the multicast routes given a certain vlanId. It prints all the traces given a certain group, source and source connect point. It has two “standard” verbosity levels.

t3-troubleshoot-mcast Help

onos> t3-troubleshoot-mcast --help
DESCRIPTION
        onos:t3-troubleshoot-mcast

    Traces all the mcast routes present in the system

SYNTAX
        onos:t3-troubleshoot-mcast [options]

OPTIONS
        -j, --json
                Output JSON
        --help
                Display this help message
        -vv, --veryverbose
                Outputs details of every trace
        -vid, --vlanId
                Vlan of incoming packet
                (defaults to None)
        -v, --verbose
                Outputs trace for each host to host combination

REST API

T3 now also offers a way to execute the troubleshoot through REST API endpoints. You can find the description and test the API at the ONOS docs endpoint here:

<ONOS_IP>:8181/onos/v1/docs/#/t3

The rest API offered are a non verbose and a verbose version of the troubleshoot simple command.

It bases itself on top of some configuration present in ONOS through netcfg, namely the VLAN interface configuration and the segment routing leaf router MAC for L3 Unicast.

The use need to specify just the Source Host, the destination host and the ETH type.

curl -X GET --header 'Accept: application/json' 'http://localhost:8181/onos/v1/t3/t3/simple/<host_src>/<host_dst>/<eht_type>'

curl -X GET --header 'Accept: application/json' 'http://localhost:8181/onos/v1/t3/t3/simple/<host_src>/<host_dst>/<eht_type>/verbose

Last Resort

In the case where you can’t figure out what is going wrong even with help of T3, you can seek help on Trellis developer mailing list trellis-dev@opennetworking.org.

There are a few things we would like you to attach:

  • Issue description

  • Environment description, such as ONOS version, switch model and OF-DPA version

  • Steps of reproduction, as detail as possible

  • Diagnostics.

We have built a tool onos-diagnostics 1 2 to help you easily collect and package ONOS diagnostics.

Alternatively, you can use onos-diagnostics-k8s 3 2.

Reference

1(1,2)

https://github.com/opennetworkinglab/onos/blob/master/tools/package/runtime/bin/onos-diagnostics

2(1,2,3,4)

https://wiki.onosproject.org/display/ONOS/ONOS+Remote+Admin+Tools

3(1,2)

https://github.com/opennetworkinglab/onos/blob/master/tools/package/runtime/bin/onos-diagnostics-k8s

4(1,2,3)

https://github.com/opennetworkinglab/routing/blob/master/trellis/trellis.py