Skip to content

FRR Release 10.1
Compare
Choose a tag to compare
@Jafaral Jafaral released this 05 Aug 06:49
· 1523 commits to master since this release
frr-10.1
14c29f8

We are pleased to announce FRR release 10.1.

FRR 10.1 brings a long list of enhancements and fixes with 601 commits from 58 developers. Thanks to all contributors.

Debian Packages - https://deb.frrouting.org

RPM Packages - https://rpm.frrouting.org

Snaps - https://snapcraft.io/frr

Docker - quay.io/frrouting/frr:10.1.0

Release Overview

Breaking changes

Enable BGP dynamic capability by default for datacenter profile

Advertise BGP "Dynamic" capability by default if using a datacenter profile. The dynamic capability gives more flexibility in terms of changing some parameters (e.g. Graceful-Restart, Long-lived Graceful-Restart timers, Addpath, Role, etc.) without resetting the session.

More details here.

Split BGP rpki cache command into separate per SSH/TCP

The old command is broken at some level. When configuring a TCP session with the source, the command thinks it's an SSH session with a username.

More details here, and here.

Add deprecation cycle for OSPF router-info X [A.B.C.D] command

Features

BGP dampening per-neighbor support

It is now possible to configure BGP dampening parameters on a per-neighbor basis. In previous releases, BGP dampening could only be configured globally or per-SAFI.

More details here.

BMP send-experimental stats

We added an option to send experimental BMP (RFC 7854) stats [65531-65534].

RFC 7854 defines BMP statistics types:

Values 65531 through 65534 are Experimental, and value 65535 is Reserved.

More details here.

Implement extended link-bandwidth for BGP

By default bandwidth in extended communities is encoded in IEEE floating-point format, and is limited to a maximum of 25 Gbps. Since not every vendor implements this correctly (due to IEEE floating-point), another draft is implemented to encode the bandwidth into IPv6 address-specific extended community.

More details here.

Paths Limit for Multiple Paths in BGP

Implemented this draft as an extension for the Addpath capability, that tells the sender to send only an arbitrary number of paths per prefix instead of sending all of the known paths.

More details here.

New command for OSPFv2 ip ospf neighbor-filter NAME [A.B.C.D]

Configure an IP prefix list to filter packets received from OSPF neighbors on the OSPF interface.

More details here.

Implement non-broadcast support for point-to-multipoint networks

This extends non-broadcast support to point-to-multipoint networks.

The AllOSPFRouters (224.0.0.5) is still joined for non-broadcast networks since it is joined for NBMA networks.

More details here.

Other significant changes

bgpd

  • Fix route leaking from the default l3vrf
  • Fix match peer when switching between IPv4/IPv6/interface
  • Fix dynamic peer graceful restart race condition
  • Fix colored routes not installed after a switchover
  • Fix crash when deleting the SRv6 locator
  • Fix no set as-path prepend ASNUM...
  • Fix negative commands for Graceful-Restart operations (avoid entering incorrect state)
  • Fix ipv4-mapped ipv6 on non 6pe
  • Fix show run of network route-distinguisher
  • Fix display when using missing-as-worst
  • Fix show bgp neighbors output
  • Fix error handling for MP/GR capabilities as a dynamic capability
  • Fix error handling when receiving BGP Prefix-SID attribute
  • Fix route-target display with a dotted format
  • Fix no bgp as-path access-list
  • Fix no form for neighbor X capability software-version
  • Check against extended community unit size for link bandwidth
  • Make sure we have enough data to handle extended link bandwidth
  • Check if FQDN capability length is in valid ranges
  • Allow using different ASNs per VRF instances
  • Send End-of-RIB not only if Graceful-Restart capability is received
  • Implement backpressure to avoid CPU hog
  • Ignore validating the attribute flags if path-attribute is configured
  • Prevent deletion of BGP peer groups associated with bgp listen range
  • Inherit some peer flags from the peer-group
  • Allow specification of AS 0 for RPKI commands
  • Allow using maximum-prefix for EVPN
  • Increase install/uninstall speed of EVPN VNIs
  • Update default-originate route-map actual map structure
  • Include unsuppress-map as a valid outgoing eBGP policy
  • Allow dynamically disable graceful-restart/long-lived graceful-restart
  • Unset advertised capabilities if the capability is disabled
  • Aggregated summary-only remove suppressed from EVPN

isisd

  • Fix crash when deactivating ISIS adjacency on the interface
  • Fix show isis database [detail] json
  • Fix show isis algorithm
  • Fix crash when configuring the circuit type for the interface
  • Fix IP/IPv6 reachability TLVs
  • When the metric-type is configured as "wide", the IS-IS generates incorrect metric values for IPv4 directly connected routes
  • Add link state support for SRv6 adjacencies
  • The hold time of hello packets on a P2P link does not match the sending interval

mgmtd

  • Implement YANG RPC/action support

ospfd

  • Fix crash in OSPF TE parsing
  • Fix the bug where ip_ospf_dead-interval_minimal_hello-multiplier did not reset the hello timer
  • Fix no write-multiplier command
  • Fix no maximum-paths command
  • Solved crash in RI parsing with OSPF TE
  • Assure OSPF AS External routes are installed after the link flap
  • Send LS Updates in response to LS Request as unicast

ospf6d

  • Handle topo change in Graceful-Restart Helper mode for max-age LSAs
  • Prevent heap-buffer-overflow with an unknown type
  • Redistribute metric for AS-external route
  • Fix next-hop computation for inter-area multi-ABR ECMP
  • Fix interface type vs. connected routes updates

pathd

  • Retry synchronous label-manager ZAPI connection

pimd

  • Fix null register before aging out reg-stop
  • Fix dr-priority range
  • Fix crash unconfiguring rp keepalive timer

lib

  • Fix keychain NB crash
  • Do not convert EVPN prefixes into IPv4/IPv6 if not needed

ripd

  • Fix clear ip rip command

ripngd

  • Fix clear ipv6 ripng command

tools

  • Handle seq num for BGP as-path in frr-reload.py

vtysh

  • Fix 'show ip[v6] prefix-list ... json' formatting by moving it to vtysh
  • Fix show route-map command when calling via do
  • Show ip ospf network ... even if it's not the same as the interface type

zebra

  • Fix mpls label bind command
  • Fix excessive exit commands
  • Fix static SRv6 segment-list SID order
  • Fix JSON output for show route summary json
  • Fix malformed json output for multiple vrfs in command show ip route vrf all json
  • Fix crash if MAC-VLAN link in another netns
  • Fix crash on MAC-VLAN link down/up
  • Deny the routes if ip protocol CLI refers to an undefined route-map
  • Bridge flap handle VLAN membership update
  • Add show fpm status [json] command

Full Changelog: frr-10.0...frr-10.1

Assets 2
Loading