TIPC-LINK(8)

okay bestie, gather round, because this is the manual page for tipc-link from the Linux manual set, and it is genuinely so much less scary than it looks. it shows you links and lets you mess with link properties. that's the whole vibe. ๐Ÿฆ‹


โœจ SYNOPSIS โœจ

tipc link set
        [ { priority PRIORITY | tolerance TOLERANCE | window WINDOW }
        link LINK ] |
        [ { broadcast [ BROADCAST  | REPLICAST  | AUTOSELECT [ ratio
        SIZE ] ] } ]

tipc link get
        [ { priority | tolerance | window } link LINK ] |
        [ { broadcast } ]

tipc link statistics { show [ link LINK ] | reset link LINK }

tipc link list

tipc link monitor set { threshold }

tipc link monitor get { threshold }

tipc link monitor summary

tipc link monitor list
        [ media  { eth | ib } device DEVICE ] |
        [ media udp name NAME ]

like yes it's a lot of brackets. breathe. it's fine. ๐Ÿ’…


๐ŸŽ€ OPTIONS ๐ŸŽ€

these are the flags (options) you can drop anywhere in the command chain. not just at the end. anywhere. iconic behavior honestly.

  • -h, --help โ€” shows help about the last valid command. for example, tipc link --help shows link help, and tipc --help shows general help. where you put the option in the string literally does not matter. ๐Ÿ’ซ
  • -j, -json โ€” spits the results out in JavaScript Object Notation (JSON).
  • -p, -pretty โ€” the default JSON format is compact and more efficient to parse, but ngl it's hard for most people to read. this flag adds indentation so it's actually readable. ๐Ÿซถ

๐Ÿ’œ DESCRIPTION ๐Ÿ’œ

  • ACTIVE link state โ€” an ACTIVE link is out here serving traffic. two links to the same node can become ACTIVE if they have the same link priority. and if there's more than two links with the same priority, the extra links get put in STANDBY state. sorry bestie, only so many can be main characters. ๐Ÿ‘‘
  • STANDBY link state โ€” a STANDBY link has lower link priority than an ACTIVE link. it's got control traffic flowing and is ready to take over should the ACTIVE link(s) go down. it's giving understudy energy and honestly we love her for it. ๐ŸŽญ
  • MTU โ€” the Maximum Transmission Unit. the two endpoints advertise their default or configured MTU at initial link setup, and they'll agree to use the lower of the two values if they differ.
  • Packets โ€” the total amount of transmitted or received TIPC packets on a link. this includes fragmented and bundled packets.
  • Fragments โ€” shown in the form fragments/fragmented. fragmented is the amount of data messages that have been broken into fragments, and then fragments is the total amount of packets that those fragmented messages got broken into.
  • Bundles โ€” shown in the form bundles/bundled. if a link becomes congested, the link will attempt to bundle data from small bundled packets into bundles of full MTU size packets before they're transmitted. โœจ
  • Profile โ€” shows the average packet size in octets/bytes for a sample of packets. it also shows the packet size distribution of the sampled packets in these intervals:

0-64 bytes 64-256 bytes 256-1024 bytes 1024-4096 bytes 4096-16384 bytes 16384-32768 bytes 32768-66000 bytes

  • Message counters
    • states โ€” Number of link state messages
    • probes โ€” Link state messages with probe flag set. typically sent when a link is idle
    • nacks โ€” Number of negative acknowledgment (NACK) packets sent and received by the link
    • defs โ€” Number of packets received out of order
    • dups โ€” Number of duplicate packets received
  • Congestion link โ€” the number of times an application has tried to send data when the TIPC link was congested.
  • Send queue โ€” Max is the maximum amount of messages that has resided in the out queue during the statistics collection period of a link. Avg is the average outqueue size during the lifetime of a link.
  • priority โ€” the priority between logical TIPC links to a particular node. link priority can range from 0 (lowest) to 31 (highest).
  • tolerance โ€” link tolerance specifies the maximum time in milliseconds that TIPC will allow a communication problem to exist before taking the link down. the default value is 1500 milliseconds.
  • window โ€” the link window controls how many unacknowledged messages a link endpoint can have in its transmit queue before TIPC's congestion control mechanism is activated.

๐Ÿงš Monitor properties

  • threshold โ€” the threshold specifies the cluster size exceeding which the link monitoring algorithm will switch from "full-mesh" to "overlapping-ring". if set to 0 the overlapping-ring monitoring is always on, and if set to a value larger than the anticipated cluster size, the overlapping-ring is disabled. the default value is 32.

๐Ÿ” Monitor information

  • table_generation โ€” represents the event count in a node's local monitoring list. it steps every time something changes in the local monitor list, including changes in the local domain.
  • cluster_size โ€” represents the current count of cluster members.
  • algorithm โ€” the current supervision algorithm used for neighbour monitoring for the bearer. possible values are full-mesh or overlapping-ring.
  • status โ€” the node status derived by the local node. possible statuses are up or down.
  • monitored โ€” represents the type of monitoring chosen by the local node. possible values are direct or indirect.
  • generation โ€” represents the domain generation, which is the event count in a node's local domain. every time something changes (peer add/remove/up/down) the domain generation is stepped and a new version of node record is sent to inform the neighbors about this change. the domain generation helps the receiver of a domain record know whether it should ignore or process the record. ๐Ÿ’ญ
  • applied_node_status โ€” the node status reported by the peer node for the succeeding peers in the node list. the node list is a circular list of ascending addresses starting with the local node. possible statuses are: U or D. the status U implies up and D down.
  • [non_applied_node:status] โ€” represents the nodes and their status as reported by the peer node. these nodes were not applied to the monitoring list for this peer node. they're usually transient and occur during the cluster startup phase or network reconfiguration. possible statuses are: U or D. the status U implies up and D down.

๐Ÿ“ก Broadcast properties

  • BROADCAST โ€” forces all multicast traffic to be transmitted via broadcast only, irrespective of cluster size and number of destinations.
  • REPLICAST โ€” forces all multicast traffic to be transmitted via replicast only, irrespective of cluster size and number of destinations.
  • AUTOSELECT โ€” auto switching to broadcast or replicast depending on cluster size and destination node number.
  • ratio SIZE โ€” set the AUTOSELECT criteria, percentage of destination nodes vs cluster size.

๐ŸŒˆ EXAMPLES ๐ŸŒˆ

tipc link monitor list

Shows the link monitoring information for cluster members on device data0. ๐Ÿ”ฅ

tipc link monitor summary

The monitor summary command prints the basic attributes. cute and to the point, periodt. โœจ


๐Ÿ’€ EXIT STATUS ๐Ÿ’€

Exit status is 0 if command was successful, or a positive integer upon failure. so if you see a number bigger than zero, something went wrong bestie, no cap.


๐Ÿซถ SEE ALSO ๐Ÿซถ

tipc(8), tipc-media(8), tipc-bearer(8), tipc-nametable(8), tipc-node(8), tipc-peer(8), tipc-socket(8)


๐Ÿ› REPORTING BUGS ๐Ÿ›

Report any bugs to the Network Developers mailing list <netdev@vger.kernel.org>, where the development and maintenance is primarily done. you do not have to be subscribed to the list to send a message there. so like, don't be shy. ๐Ÿ™ˆ


๐Ÿ‘‘ AUTHOR ๐Ÿ‘‘

Richard Alpe <richard.alpe@ericsson.com>


and that's the whole page, bestie!! this one comes from the iproute2 crew, dated 22 Mar 2019 โ€” go forth and monitor those links, you absolute icon ๐Ÿ’–โœจ๐ŸŽ€

๐Ÿ“ฆ where this page comes from