- frog - lists.frrouting.org

MAC Flapping With VRRP on Linux
by Alasdair Muckart 19 May '24

19 May '24

Hello FRR folk, I'm having difficulty with traffic to linux (Ubuntu 22.04, kernel 5.15) routers running VRRP. The problem is MAC flapping between the VRRP MAC and the underlying interface MAC. It's so bad traffic from the LAN to the VIP is basically unusable. Every who-has for the VIP or the VRRP primary's underlying interface IP gets multiple responses, and the mac table on the switches is flailing. I've tried all the combinations of the various arp sysctl I can think of and I can't get one that will only respond to requests for the VIP with the VRRP MAC. Either I get duelling replies with both the VIP MAC and the underlying interface MAC, or I get nothing at all. Can anyone tell me what I need to do to get the routers to only reply with the VIP MAC when there's an arp who-has for the VIP? I couldn't see anything in the manual about this. TIA. In case it's relevant, the topology as follows: A pair of core switches connected by an ERPS ring. Two routers, each connected to both switches with an active/passive bond interface. VRRP running on the bond interface. The bond interfaces are .2 and .3, the VIP is .1. The eth0 and eth1 interfaces are unnumbered children of the bond0. 192.168.1.2/24 eth0 eth1 +-----------+bond0+-X----------+ | | | | | +--+-----+--+ | | | vrrp4-1-1 | | | +-----------+ | | 192.168.1.1/24 | | | +------+-----+ +------+-----+ | +-----------------+ | | SWITCH 1 | ERPS | SWITCH 2 | | +---------------X-+ | +------+-----+ +------+-----+ | | | | | +-----------+ | | | vrrp4-1-1 | | | +--+-----+--+ | | | | | +-----------+bond0+-X----------+ eth0 eth1 192.168.1.3/24 I've got the bond0 interfaces because FRR doesn't seem to cope at all with having two interfaces in the same VRRP on the same router, one of them is permanently stuck 'initializing', but the MAC flapping is the same with just one interface and no bond. -- Alasdair Muckart (he/him) Network Infrastructure Architect Catalyst.Net Limited - Expert Open Source Solutions Catalyst.Net Ltd - a Catalyst IT group company DDI: +64 4 897 7794 | Mobile: +64 22 638 5141 | Tel: +64 4 499 2267 | www.catalyst.net.nz CONFIDENTIALITY NOTICE: This email is intended for the named recipients only. It may contain privileged, confidential or copyright information. If you are not the named recipient, any use, reliance upon, disclosure or copying of this email or its attachments is unauthorised. If you have received this email in error, please reply via email or call +64 4 499 2267.

1 0

/etc vs /usr/local/etc
by Randy Bush 16 May '24

16 May '24

frr n00b here setting up frr as a debian12 amd64 kvm image. and it was working from the packaged binary. cool beans. except i run is-is, which means i had to build from source (which jammed on `rtrlib` needing a version of `libssh` i still need to find. but i disabled rpki <blush>), and completed the build and install on branch `stable/10.0`. it blows chunks on start cannot run start: /usr/local/etc/frr/daemons does not exist the binary package for deb12 put configs in `/etc/frr`, not `/usr/local/etc/frr`, and that's kinda the debian way. (yes, i used to love freebsd, but ...) a bit of poking finds that it installed and expects `/usr/local/...` for everything. and i see no `./configure` option to set `PREFIX`. so i hack `./configure` - ac_default_prefix=/usr/local + ac_default_prefix= it builds and installs, but then i find myself at the top of what looks like a much deeper rabbit hole. frr.service: Failed to locate executable /usr/local/sbin/frrinit.sh: No such file or directory clue bat appreciated. randy

2 5

FRR support for "IPv6 over SRv6 Core"
by Balazs Varga 16 May '24

16 May '24

Hi, FRR seems to support only “Transposition Scheme” of RFC9252. Sure, that is enough for L3VPN, but not an option for “vrf default” i.e., the global routing table. Does FRR support RFC9252, Section 5.4. Global IPv6 over SRv6 Core? Many thanks Bala'zs

1 0

IS-IS unnumbered support for IPv4?
by Hendrik Visage 15 May '24

15 May '24

Trying to use (similar to OpenFabric) IS-IS with unnumbered interfaces. The problem is that the IPv6 loopbacks are propagated, but the IPv4 loopbacks aren't, contrary to Openfabric, without IPs on the interfaces see below I've searched high and low, and couldn't find any setting I've missed, and the source code haven't been too help full to see/understand if I'm missing something or if the needed haven't been coded into IS-IS yet. ``` Current configuration: ! frr version 10.0 frr defaults datacenter hostname margherita log syslog informational service integrated-vtysh-config ! interface enp67s0f1 ip router isis 1 ipv6 router isis 1 isis csnp-interval 2 isis hello-interval 1 isis hello-multiplier 2 isis network point-to-point exit ! interface enp67s0f2 ip router isis 1 ipv6 router isis 1 isis csnp-interval 2 isis hello-interval 1 isis hello-multiplier 2 isis network point-to-point exit ! interface enp67s0f3 ip router isis 1 ipv6 router isis 1 isis csnp-interval 2 isis hello-interval 1 isis hello-multiplier 2 isis network point-to-point exit ! interface lo ip address 10.4.127.1/32 ip router isis 1 ipv6 address 2c0f:dead:beef:feed:2:0:4:1/128 ipv6 router isis 1 isis passive exit ! router isis 1 is-type level-1 net 49.1004.feed.feed.0004.0001.00 lsp-timers gen-interval 1 refresh-interval 180 max-lifetime 600 redistribute ipv4 connected level-1 redistribute ipv6 connected level-1 topology ipv4-multicast topology ipv4-mgmt topology ipv6-unicast topology ipv6-multicast topology ipv6-mgmt topology ipv6-dstsrc log-adjacency-changes exit ! end ``` ``` root@margherita:~# vtysh -c "show isis route" Area 1: IS-IS paths to level-1 routers that speak IP Vertex Type Metric Next-Hop Interface Parent margherita 10.4.127.1/32 IP internal 0 margherita(4) pu.bl.ic1.0/24 IP internal 0 margherita(4) IS-IS L1 IPv4 routing table: Prefix Metric Interface Nexthop Label(s) ------------------------------------------------------- 10.4.127.1/32 0 - - - pu.bl.ic1.0/24 0 - - - IS-IS paths to level-1 routers that speak IPv6 Vertex Type Metric Next-Hop Interface Parent margherita 2c0f:dead:beef:feed:2:0:4:1/128 IP6 internal 0 margherita(4) quesadilla TE-IS 10 quesadilla enp67s0f1 margherita(4) guacamole TE-IS 10 guacamole enp67s0f2 margherita(4) tequila TE-IS 10 tequila enp67s0f3 margherita(4) 2c0f:dead:beef:feed:2:0:4:2/128 IP6 internal 10 quesadilla enp67s0f1 quesadilla(4) 2c0f:dead:beef:feed:2:0:4:3/128 IP6 internal 10 guacamole enp67s0f2 guacamole(4) 2c0f:dead:beef:feed:2:0:4:4/128 IP6 internal 10 tequila enp67s0f3 tequila(4) IS-IS L1 IPv6 routing table: Prefix Metric Interface Nexthop Label(s) ---------------------------------------------------------------------------------------- 2c0f:dead:beef:feed:2:0:4:1/128 0 - - - 2c0f:dead:beef:feed:2:0:4:2/128 10 enp67s0f1 fe80::7777:55ff:fe31:6648 - 2c0f:dead:beef:feed:2:0:4:3/128 10 enp67s0f2 fe80::7777:55ff:fe31:6664 - 2c0f:dead:beef:feed:2:0:4:4/128 10 enp67s0f3 fe80::7777:55ff:fe31:663c - IS-IS paths to level-1 routers that support IPv6 dst-src routing Vertex Type Metric Next-Hop Interface Parent margherita quesadilla TE-IS 10 quesadilla enp67s0f1 margherita(4) guacamole TE-IS 10 guacamole enp67s0f2 margherita(4) tequila TE-IS 10 tequila enp67s0f3 margherita(4) IS-IS L1 IPv6 (dst-src routing) routing table: ``` adding the "loopback" IP to all connected interfaces: ``` root@margherita:~# for i in 0 1 2 3; do ip a add dev enp67s0f$i 10.4.127.1 ;done root@margherita:~# vtysh -c "show isis route" Area 1: IS-IS paths to level-1 routers that speak IP Vertex Type Metric Next-Hop Interface Parent margherita 10.4.127.1/32 IP internal 0 margherita(4) pu.bli.c1.0/24 IP internal 0 margherita(4) quesadilla TE-IS 10 quesadilla enp67s0f1 margherita(4) guacamole TE-IS 10 guacamole enp67s0f2 margherita(4) tequila TE-IS 10 tequila enp67s0f3 margherita(4) 10.4.127.2/32 IP TE 10 quesadilla enp67s0f1 quesadilla(4) pu.bli.c1.0/24 IP TE 10 quesadilla enp67s0f1 quesadilla(4) guacamole enp67s0f2 guacamole(4) tequila enp67s0f3 tequila(4) 10.4.127.3/32 IP TE 10 guacamole enp67s0f2 guacamole(4) 10.4.127.4/32 IP TE 10 tequila enp67s0f3 tequila(4) IS-IS L1 IPv4 routing table: Prefix Metric Interface Nexthop Label(s) ---------------------------------------------------------- 10.4.127.1/32 0 - - - 10.4.127.2/32 10 enp67s0f1 10.4.127.2 - 10.4.127.3/32 10 enp67s0f2 10.4.127.3 - 10.4.127.4/32 10 enp67s0f3 10.4.127.4 - pu.bli.c1.0/24 10 enp67s0f1 10.4.127.2 - enp67s0f2 10.4.127.3 - enp67s0f3 10.4.127.4 - IS-IS paths to level-1 routers that speak IPv6 Vertex Type Metric Next-Hop Interface Parent margherita 2c0f:dead:beef:feed:2:0:4:1/128 IP6 internal 0 margherita(4) quesadilla TE-IS 10 quesadilla enp67s0f1 margherita(4) guacamole TE-IS 10 guacamole enp67s0f2 margherita(4) tequila TE-IS 10 tequila enp67s0f3 margherita(4) 2c0f:dead:beef:feed:2:0:4:2/128 IP6 internal 10 quesadilla enp67s0f1 quesadilla(4) 2c0f:dead:beef:feed:2:0:4:3/128 IP6 internal 10 guacamole enp67s0f2 guacamole(4) 2c0f:dead:beef:feed:2:0:4:4/128 IP6 internal 10 tequila enp67s0f3 tequila(4) IS-IS L1 IPv6 routing table: Prefix Metric Interface Nexthop Label(s) ---------------------------------------------------------------------------------------- 2c0f:dead:beef:feed:2:0:4:1/128 0 - - - 2c0f:dead:beef:feed:2:0:4:2/128 10 enp67s0f1 fe80::7777:55ff:fe31:6648 - 2c0f:dead:beef:feed:2:0:4:3/128 10 enp67s0f2 fe80::7777:55ff:fe31:6664 - 2c0f:dead:beef:feed:2:0:4:4/128 10 enp67s0f3 fe80::7777:55ff:fe31:663c - IS-IS paths to level-1 routers that support IPv6 dst-src routing Vertex Type Metric Next-Hop Interface Parent margherita quesadilla TE-IS 10 quesadilla enp67s0f1 margherita(4) guacamole TE-IS 10 guacamole enp67s0f2 margherita(4) tequila TE-IS 10 tequila enp67s0f3 margherita(4) IS-IS L1 IPv6 (dst-src routing) routing table: root@margherita:~# ```

1 0

Isn't ospf ban an ABR to flood lsa3 to a non-backbone area?
by chan alfie 09 May '24

09 May '24

hi community, I got some confused of ospf lsa3. topology: h1 -- r1 --a12-- r2 --a0-- r3 --a34-- r4 -- h4 | | +---a23 ---+ I am using mininet to emulate the above environment. Here are the topo desc: area 12: r1(intf:r1r2) and r2(intf:r2r1) , connected network: 10.0.12.0/30 area 0: r2(intf:r20r30) and r3(intf:r30r20), connected network: 10.0.23.0/30 area 23: r2(intf:r2r3) and r3(intf:r3r2), connected network: 10.0.23.4/30 area 34: r3(intf:r3r4) and r4(intf:r4r3), connected network: 10.0.34.0/30 r1 has a stub net: 10.0.1.0/254, connected to a host, h1 r4 has a stub net: 10.0.4.0/254, connected to a host, h4 my question is: why r2/r3 generated lsa3 and advertise on area 23, isn't it prohibited ABR to flood lsa3 to non-backbone area? r2 ``` # show ip ospf database Summary Link States (Area 0.0.0.23) Link ID ADV Router Age Seq# CkSum Route 10.0.0.1 10.0.0.2 448 0x80000001 0xdc5f 10.0.0.1/32 10.0.0.1 10.0.0.3 449 0x80000001 0x3bf5 10.0.0.1/32 10.0.0.2 10.0.0.2 498 0x80000001 0x6ed6 10.0.0.2/32 10.0.0.2 10.0.0.3 449 0x80000001 0xcc6d 10.0.0.2/32 10.0.0.3 10.0.0.2 448 0x80000001 0xc871 10.0.0.3/32 10.0.0.3 10.0.0.3 498 0x80000001 0x5ee4 10.0.0.3/32 10.0.0.4 10.0.0.2 128 0x80000001 0x230c 10.0.0.4/32 10.0.0.4 10.0.0.3 129 0x80000001 0xb87f 10.0.0.4/32 10.0.1.0 10.0.0.2 448 0x80000001 0x40f1 10.0.1.0/24 10.0.1.0 10.0.0.3 449 0x80000001 0x9e88 10.0.1.0/24 10.0.4.0 10.0.0.2 128 0x80000001 0x83a1 10.0.4.0/24 10.0.4.0 10.0.0.3 129 0x80000001 0x1915 10.0.4.0/24 10.0.12.0 10.0.0.2 498 0x80000001 0x50e3 10.0.12.0/30 10.0.12.0 10.0.0.3 449 0x80000001 0xae7a 10.0.12.0/30 10.0.23.0 10.0.0.2 498 0x80000001 0xd652 10.0.23.0/30 10.0.23.0 10.0.0.3 498 0x80000001 0xd057 10.0.23.0/30 10.0.34.0 10.0.0.2 448 0x80000001 0xc152 10.0.34.0/30 10.0.34.0 10.0.0.3 498 0x80000001 0x57c5 10.0.34.0/30 ``` any idea?

2 2

Slack invites?
by Hendrik Visage 07 May '24

07 May '24

Good day, The invite system seems to be limiting access only to a set of domains, and requesting access, seems to have a continuous loop only allowing certain domains and telling to contact the slack admin?

5 5

ospf? invalid link address
by Michael DiDomenico 30 Apr '24

30 Apr '24

my frr log seems to be spewing the following lines zebra: extended error: invalid link address netlink-dp error: invalid argument, type=rtm_newneigh(28), seq=..., pid=... nl_batch_read_resp: netlink error message seq=... nlmsghdr [len=64 type(2) error flags=(0x0300) {dump,(root|replace|capped),(match|exclude|ack_tlvs)} seq=... pid=...] can anyone suggest some debug settings that i can turn on to help figure out what's going on. i'm not entirely sure when this started, i only noticed it because my logs dir filled up

1 1

BGP community regex rewrite
by Piotr Ginalski 21 Apr '24

21 Apr '24

Hi Is it possible to use regex to rewrite the BGP communities? Let's say I have extended community 65000:[0-9]+ and route-map which matches it. Is it possible to get the regex value and supply it to set community 0:$1 etc.? Cheers Piotr Ginalski

1 0

Migration scripts for FRR from Quagga
by Doug Tucker 17 Apr '24

17 Apr '24

I am needing to undertake migrating from all of my quagga instances to frr. I have read through the general documentation but before I started from scratch I thought I would see if there were some migration/conversion scripts available for existing quagga installations. Thank you for your assistance. Doug Tucker

1 0

IS-IS on unnumbered interfaces?
by Hendrik Visage 17 Apr '24

17 Apr '24

Good day, Busy FAFOing on a new setup to implement EVPN on ProxMox 8.1+, and as OpenFabric failed on the links with multiple nodes, (the network I have for redundancy at present) I tried IS-IS, which "worked", except I've ran into the unnumbered problem w.r.t. IPv4 (IPv6 is happily exchanging routes and update routing tables) when I have the direct connect ethernet interfaces without IPs. I see BGP/OSPF unnumbered stuff, but nothing yet about IS-IS for IPv4, or how to set up IP unnumbered stuff using FRR on Linux. How is this (usually) handled by others doing IS-IS ?

1 0