- Bird-users - Mailing Lists

Way too many routes imported in OSPF
by Alarig Le Lay 29 Jan '19

29 Jan '19

Hi, I’ve set up an OSPF as an IGP to resolve the BGP next-hops on the routes learned from eBGP sessions. On one of my border routers (running FreeBSD), the counters are correct: root@nominoe:~ # birdc 'show protocols all "ospf_*"' BIRD 2.0.2 ready. Name Proto Table State Since Info ospf_ipv4 OSPF master4 up 16:29:14.755 Running Channel ipv4 State: UP Table: master4 Preference: 150 Input filter: (unnamed) Output filter: (unnamed) Import limit: 200 Action: block Routes: 59 imported, 0 filtered, 26 exported Route change stats: received rejected filtered ignored accepted Import updates: 114 0 0 0 114 Import withdraws: 43 0 --- 0 43 Export updates: 4343089 79 4342789 --- 221 Export withdraws: 25677 --- --- --- 195 ospf_ipv6 OSPF master6 up 16:29:14.755 Running Channel ipv6 State: UP Table: master6 Preference: 150 Input filter: (unnamed) Output filter: (unnamed) Import limit: 200 Action: block Routes: 16 imported, 0 filtered, 13 exported Route change stats: received rejected filtered ignored accepted Import updates: 178 0 0 0 178 Import withdraws: 161 0 --- 0 161 Export updates: 210288 168 210107 --- 13 Export withdraws: 4415 --- --- --- 0 On the other (running Gentoo), I can’t have so much: asbr02 ~ # birdc 'show protocols all "ospf_*"' BIRD 2.0.3 ready. Name Proto Table State Since Info ospf_ipv4 OSPF master4 up 17:07:45.751 Running Channel ipv4 State: UP Table: master4 Preference: 150 Input filter: (unnamed) Output filter: (unnamed) Import limit: 200 Action: block Routes: 64 imported, 0 filtered, 4294120544 exported Route change stats: received rejected filtered ignored accepted Import updates: 258 0 0 0 258 Import withdraws: 181 0 --- 0 181 Export updates: 2107116 242 2106865 --- 9 Export withdraws: 21844 --- --- --- 846761 ospf_ipv6 OSPF master6 up 17:02:14.667 Running Channel ipv6 State: UP Table: master6 Preference: 150 Input filter: (unnamed) Output filter: (unnamed) Import limit: 200 Action: block Routes: 19 imported, 0 filtered, 4294870086 exported Route change stats: received rejected filtered ignored accepted Import updates: 204 0 0 0 204 Import withdraws: 185 0 --- 0 185 Export updates: 325977 194 325773 --- 10 Export withdraws: 10926 --- --- --- 97220 asbr02 ~ # birdc 'show route count table master4' BIRD 2.0.3 ready. 2461149 of 2461149 routes for 737465 networks in table master4 asbr02 ~ # birdc 'show route count table master6' BIRD 2.0.3 ready. 239266 of 239266 routes for 63527 networks in table master6 I already tried to restart the protocols, but it doesn’t change anything. Any idea? Thanks, -- Alarig

3 7

Default config file name 1.6 vs. 2.0
by Eugene Crosser 28 Jan '19

28 Jan '19

Hello, I would like to suggest to change the default name of the configuration file for bird version 2.0. The two bird versions cannot use the same file because of syntax change. And some of us distribute the file via configuration management tools (like ansible etc.). This makes upgrade path rather difficult and unnecessarily error-prone: you need to synchronize exchange of the version of the configuration file with upgrade (or downgrade! That can be necessary too!) of the bird package. Whereas if different versions used different files, you could distribute both files with you configuration management tool in advance, and then gradually upgrade the package, with easy rollback if things do not go well. Even if configuration management is not used, it is still more convenient to be able to start one or the other version of the daemon without the need to specify alternative configuration, or replacing the file. Thanks for consideration! Eugene

4 5

Exporting multipath link routes from Linux kernel
by Eugene Crosser 28 Jan '19

28 Jan '19

Hello all, we do virtual hosting, and we provide routeable /32 addresses to the guests. Kernel routes on the KVM host are link routes that look like this: 1.0.0.113 dev pub020304050612 proto static They are picked up by bird and exported to the core router. When we launch multiple guests with the same address on multiple hosts, this results in an ECMP route on the core router, providing load balancing. This all works fine until we launch more than one guest with the same address on _one_ host. We create kernel multipath route that looks like this: 1.0.0.115 proto static metric 10 nexthop dev pub020304050612 weight 1 nexthop dev pub020304050616 weight 1 Note that there is no "via" address in the hop configulations! This actually works, i.e. connections originating from the host are balanced between those guests. But bird refuses to pick up such route because of the code here: https://gitlab.labs.nic.cz/labs/bird/blob/master/sysdep/linux/netlink.c#L528 Questions: 1. What was the justification for disallowing gateway-less multipath routes? Would it make sense to allow them (in the mainstream code)? 2. Would it be sufficient to simply drop the check for the presence of the gateway address in the message, and return `first` even if gateway address was not present? Thank you, Eugene

2 5

reorder BGP plink preference
by Eugene M. Zheganin 27 Jan '19

27 Jan '19

Hello, I have a bird router serving multiple BGP sessions with my AS uplinks, and now I need to create a more complicated thing, - I have several FIBs and I need to reorder the BGP link preference for a given FIB. Say, I have 3 uplinks, A/B/C, A is the most preferable one for default FIB, but I need B to be the most preferable one for FIB 1. I have some basic experience with multiple routing tables in bird, I'm successfully using two at the moment, but I just have no idea on how to do this. I mean I know how to make the B upstream the only channel for FIB 1 (without redundancy), but I have no idea on how to reorder the preference in order to keep the redundancy. Is there such a way ? Thanks. Eugene.

2 2

A few suggestions for docs improvement
by Igor Podlesny 26 Jan '19

26 Jan '19

First of all, protocols description are better kept in separate HTML pages otherwise it's hard to find by some keywords that several protocols have. It's really inconvenient to loose focus when you search something related to OSPF and Find scrolls page to completely different protocol. Second: why is font so tiny? -- End of message. Next message?

3 2

OSPF transmit delay is not documented?
by Igor Podlesny 26 Jan '19

26 Jan '19

Whereas it's given in some default Bird configs, I couldn't find it in manual. -- End of message. Next message?

2 1

Bird OSPF bug? (Mutliple areas, rfc1583compat off, cost ignored)
by Igor Podlesny 26 Jan '19

26 Jan '19

I have test lab -- two OSPF routers, Bird 1.6.4: A, B. 2 areas defined: 0.0.0.0, 0.0.0.1. Export "all". When rfc1583compat is Off I observe default route gets propagated via area 0.0.0.1 in despite its interfaces have higher costs than area 0.0.0.0's have. Turning rfc1583compat On changes it to expected behaviour: interfaces costs are respected again. Bug or am I missing some piece of OSPF knowledge here? -- End of message. Next message?

2 3

OSPF, BSD: every additional IP on an interface gives another neighbour
by Igor Podlesny 26 Jan '19

26 Jan '19

The docs say: "... secondary switch On BSD systems, older versions of BIRD supported OSPFv2 only for the primary IP address of an interface, other IP ranges on the interface were handled as stub networks. Since v1.4.1, regular operation on secondary IP addresses is supported, but disabled by default for compatibility. This option allows to enable it. The option is a transitional measure, will be removed in the next major release as the behavior will be changed. On Linux systems, the option is irrelevant, as operation on non-primary addresses is already the regular behavior. ..." Let's focus on: "... Since v1.4.1, regular operation on secondary IP addresses is supported, but disabled by default for compatibility. This option allows to enable it. ..." My observations show that what actually happens when whether this option is On or it's Off is not anything regular. Instead for every additional IP there's "virtually" additional OSPF neighbour instance. bird> show ospf ospf1: RFC1583 compatibility: disabled Stub router: No RT scheduler tick: 1 Number of areas: 1 Number of LSAs in DB: 23 Area: 0.0.0.0 (0) [BACKBONE] Stub: No NSSA: No Transit: No Number of interfaces: 3 Number of neighbors: 3 Number of adjacent neighbors: 3 -- every IP-address on interface actually handled as separate interface with that IP, that's why this weirdness with neighbour count is seen. O_o Is it so on Linux as well, or BSD implementation specific? -- End of message. Next message?

2 1

i-bgp neighbor not coming up
by Shahan Agha 25 Jan '19

25 Jan '19

Hi All, I am facing something very strange from almost a week now. I have multiple i-bgp neighbors which are working just fine. # birdc show protocols bgp_r4 BGP master up 2019-01-21 Established bgp_r3 BGP master up 2019-01-21 Established *bgp_r2 BGP master start 15:04:30 Active Socket: Connection refused* There as in issue with only one of them. It did work initially but for some reason it is now down since a few days. I have tried everything in my knowledge but have failed to bring it to work again. Below are the configurations: # R-1 # protocol bgp bgp_r1 { # disabled; debug { states, routes, filters, interfaces, events }; router id 10.10.2.1; local as 65001; neighbor 10.10.2.2 as 65001; next hop self; direct; } # R-2 # protocol bgp bgp_r2 { # disabled; debug { states, routes, filters, interfaces, events }; router id 10.10.2.2; local as 65001; neighbor 10.10.2.1 as 65001; next hop self; direct; } Any idea what can be the possible reason behind this? Thanks, Shahan

2 2

BGP route back
by Roman Romanyak 22 Jan '19

22 Jan '19

Hello Bird users, Does anyone know if there is a way to announce a BGP route back to the router it was received from. I need to do this for the following scenario. Lets say there are two routers with 2 transit links on each with full view tables and Bird server as a route-reflector. There is a need to force traffic to a specific destination via one of the ISPs, so I match the route in the import filter and set a local pref on it. But that will only make the route server announce the route with a high local pref to a second router, where the desired transit isn't directly connected. So I get the desired effect on the second router, it will learn and install a route with high local pref. The first router where the desired transit link is connected by default selects another transit link because of the shorter as-path. Here is the import filter snippet (x.y.z.0/24 is a dest route, as-path 1234 is a directly connected ISP on router-1: if source = RTS_BGP && net = x.y.z.0/24 && bgp_path.first = 1234 then { bgp_local_pref = 150; accept; } I think that bird doesn't do that because the protocol matches on the peer and on the route. Thanks!

3 8