- Bird-users - Mailing Lists

bird version 2.0.4 - peering with cisco version ios xe 16.3.5 - invalid open message
by Marco van Tol 13 Mar '19

13 Mar '19

Hi there, I'm trying to setup iBGP peering with a cisco router which is giving me an "invalid open message" that I can't seem to fix. - The same router is peering with Juniper and other bird 1.6.3 routers without issues. - The router is also doing IPv6 based peering with all the other routers, which gives the exact same scenario as with IPv4. The relevant parts of the config look as follows (actual IP and AS spaces replaced with private spaces): -----< cut here >----- router id 10.0.0.1; protocol static default_v4 { ipv4 { preference 50; }; route 0.0.0.0/0 unreachable; } filter bgp_in_v4 { if net ~ [ 10.1.0.0/28 ] then accept; reject; } template bgp type1_v4 { direct; local 10.0.0.1 as 64512; ipv4 { export where proto = "default_v4"; import filter bgp_in_v4; import keep filtered; }; } protocol bgp peer_type1a_v4 from type1_v4 { neighbor 10.0.0.2 as 64512; } -----< cut here >----- The debug messages I'm getting are these: -----< cut here >----- 2019-03-08 11:02:33.199 <TRACE> peer_type1a_v4: Incoming connection from 10.0.0.2 (port 18581) accepted 2019-03-08 11:02:33.199 <TRACE> peer_type1a_v4: Sending OPEN(ver=4,as=64512,hold=240,id=0a000001) 2019-03-08 11:02:33.200 <TRACE> peer_type1a_v4: Got OPEN(as=64512,hold=180,id=10.0.0.2) 2019-03-08 11:02:33.200 <TRACE> peer_type1a_v4: Sending KEEPALIVE 2019-03-08 11:02:33.201 <RMT> peer_type1a_v4: Received: Invalid OPEN message 2019-03-08 11:02:33.201 <TRACE> peer_type1a_v4: State changed to stop 2019-03-08 11:02:33.201 <TRACE> peer_type1a_v4: Down -----< cut here >----- My suspicion is that the ipv6 like address representation in the sent open message router id might confuse the cisco. So the opening message we're sending seems to be having the router id represented in hex form, like 32 bits of an IPv6 address. Is this correct, and is there a way to either fix or work around this? Many thanks! Marco van Tol

2 13

maybe it is a bug in nexthop__same function
by Derek Pan 12 Mar '19

12 Mar '19

Hi I think below function should be used by comparing two netxhop object with different pointer. If this function is called, the X and Y always have different pointer values, but if their values are same, they also can be treated as the same. I think the "return x== y;" should be replaced by "return 1;". Please have a look. Thank. int nexthop__same(struct nexthop *x, struct nexthop *y) { for (; x && y; x = x->next, y = y->next) { if (!ipa_equal(x->gw, y->gw) || (x->iface != y->iface) || (x->flags != y->flags) || (x->weight != y->weight) || (x->labels_orig != y->labels_orig) || (x->labels != y->labels)) return 0; for (int i = 0; i < x->labels; i++) if (x->label[i] != y->label[i]) return 0; } return x == y; }

3 3

all protocol state debug
by Eugene M. Zheganin 11 Mar '19

11 Mar '19

Hello, I want to log all protocols state changes, and I have a question: debug all { states } is valid in birdc, but why debug all { states }; is invalid in bird.conf ? Thanks. Eugene.

2 2

Backportability of bird2 debian package
by Eugene Crosser 11 Mar '19

11 Mar '19

Hi, I have a request to Ondřej Surý: debian-packaged version of bird2 that is currently on https://salsa.debian.org/debian/bird2.git is unbuildable on older distros because of required version of debhelper (cosmic and stretch have version 11) and init-system-helpers (cosmic has 1.54, stretch - 1.48). Could these dependencies be relaxed to ease backporting of the package? Thanks, Eugene

1 0

How to update from v1.6 to v2.0
by Isaac HO 11 Mar '19

11 Mar '19

Dear All, Would I need to uninstall v1.6 or could update the version by simply steps? For v1.6 un-installation, would you mind providing me the steps? or would you please providing me the upgrading steps? Thanks, Isaac

2 1

bird 2.0.4 - should "protocol rpki" be recognised?
by Marco van Tol 08 Mar '19

08 Mar '19

Hi there, I'm trying to setup a router with RPKI checks. I'm using BIRD version 2.0.4. I was checking the documentation here: https://bird.network.cz/?get_doc&v=20&f=bird-6.html#ss6.13 When I put the following in my config, I get the error message: /etc/bird.conf:32:10 syntax error, unexpected SYM Line 32 is the line with "protocol rpki" on it. This is a copy/paste from the relevant part of the config: -----< cut here >----- roa4 table r4; roa6 table r6; protocol rpki { { roa4 { table r4; }; remote "myhost.domain" transport ssh { bird private key "/root/.ssh/rpki_validator" remote public key "/root/.ssh/known_hosts" user "rpki" }; } -----< cut here >----- What I also noticed is that when I type "show ?" in birdc, it does not list the roa option. Do I need to specify certain commandline options to enable rpki perhaps? Just thinking outloud. :-) Many thanks in advance, Marco van Tol

2 2

Showing local router ID for BGP protocols
by Jakub Nowacki 08 Mar '19

08 Mar '19

Hi, I'm trying to find a command that will allow me to verify router ID being used for particular BGP session. When using 'show protocols all <BGP_PROTOCOL_NAME>' it shows verbose info, but it doesn't show Local Router ID. I know it shows 'Source address' but this is not the same. Is there a command to verify that? Kind regards -- Kuba Nowacki

3 2

Route redistribution between OSPF instances
by Yaroslav Dronskii 08 Mar '19

08 Mar '19

Hello, I need some help in understanding bird. Route redistribution between ospf instances doesn't work for me. I tried everything I could think of short of debugging source code, and narrowed it down to route type. If it's RTS_OSPF_IA, ospf_metric2 is set to LSINFINITY. How I wanted to use it: VM announces some /32 net in area 20. This net is received by vm's host into table_ospf_0 and redistributed to: 1) backbone area so other vm hosts can reach it 2) table_allroutes via a pipe. Ospf_allroutes protocol takes it from table_allroutes and announces it to everyone in area 30. Am I doing something stupid? Should it even work? Minimal config example: ipv4 table table_ospf_0; # contains host-host vm /32 ip announces ipv4 table table_allroutes; # contains all routes learned. protocol pipe pipe_ospf_0_allroutes { table table_ospf_0; peer table table_allroutes; import none ; export filter { if (net.len = 32) then accept; else reject; } ; } protocol ospf ospf_vm { ipv4 { table table_ospf_0; export all; import filter filter_from_ospf; }; # meant for host-host vm /32 ip announces # without it, best route to other host's vm will be via PE area 0.0.0.0 { interface "ospf0-br" { authentication cryptographic; password "pass1"; }; }; # meant for ip announces. area 0.0.0.20 { interface "ospf20-br" { authentication cryptographic; password "pass2"; }; }; } protocol ospf ospf_allroutes { ipv4 { table table_allroutes; import keep filtered on; export all; import filter filter_to_kernel; }; # Meant for vm's that want to accept ALL routes with per-vrf ospf tag area 0.0.0.30 { interface "ospf30-br" { authentication cryptographic; password "pass3"; }; }; } How route is seen: Imported from other vm host via ospf: bird> show route table table_ospf_0 10.20.5.135/32 all Table table_ospf_0: 10.20.5.135/32 unicast [ospf_vm 13:52:53.209] IA (150/20) [10.255.10.129] via 10.133.146.12 on ospf0-br Type: OSPF-IA univ Piped to table_allroutes: (as expected, nothing changed) bird> show route table table_allroutes 10.20.5.135/32 all Table table_allroutes: 10.20.5.135/32 unicast [ospf_vm 13:52:53.209] IA (150/20) [10.255.10.129] via 10.133.146.12 on ospf0-br Type: OSPF-IA univ In ospf_vm: # birdc show ospf state ospf_vm area 0.0.0.0 router 10.255.10.129 distance 10 ... xnetwork 10.20.5.135/32 metric 10 ... xrouter 10.20.5.135 metric 10 area 0.0.0.20 router 10.255.10.128 distance 0 xnetwork 10.20.5.135/32 metric 20 ... xrouter 10.20.5.135 metric 20 # birdc show ospf state ospf_allroutes area 0.0.0.30 router 10.255.10.128 distance 0 external 10.20.5.135/32 metric2 16777215

3 2

OSPF authentication supported for IPv6 / OSPFv3?
by Cybertinus 05 Mar '19

05 Mar '19

Hello Bird users, I'm trying to get authentication enabled on my OSPF sessions. I'm running Bird 1.6.4 on Debian 9.8. I have the following config in my ospf.conf: protocol ospf ospf1 { import filter only_loopbacks; export filter only_loopbacks; area 0.0.0.0 { interface "lo"; interface "eno1" { type pointopoint; bfd on; }; interface "eno2" { type pointopoint; bfd on; }; interface "eno3" { type pointopoint; bfd on; authentication cryptographic; password "Test123"; }; }; } I include this ospf.conf in both bird.conf and bird6.conf. For IPv4 Bird it works without an issue and the sessions (over eno3) are authenticated. For IPv6 I see the following error when I try to load the config: # birdc6 configure check BIRD 1.6.4 ready. Reading configuration from /etc/bird/bird6.conf /etc/bird/ospf.conf, line 17: Authentication not supported in OSPFv3 When I check https://bird.network.cz/?get_doc&v=16&f=bird-6.html#ss6.8 I see the following: "authentication cryptographic An authentication code is appended to every packet. The specific cryptographic algorithm is selected by option algorithm for each key. The default cryptographic algorithm for OSPFv2 keys is Keyed-MD5 and for OSPFv3 keys is HMAC-SHA-256. Passwords are not sent open via network, so this mechanism is quite secure. Packets can still be read by an attacker." So, I think it should work for IPv6 too. What am I doing wrong? Or did I hit a bug of some kind? Kind regards, Cybertinus

2 2

BGP Graceful Restart
by Derek Pan 04 Mar '19

04 Mar '19

Hi >From the post https://bird.network.cz/pipermail/bird-users/2016-December/010760.html I understood that it should use kill -9 to stop and restart BIRD process to trigger graceful restart function. But I still have a question: Suppose in one BIRD process, there are more than two BGP protocol instances and OSPF protocol instances, And there are some BGP instances without graceful restart. Once I stop the BIRD process using kill -9 and then start it, I know those BGP instances with graceful restart would work fine But the others BGP instances without graceful restart would just be restarted. But this is not my expect behavior (I expect the others BGP and OSPF protocols would not be touched). Is there any command to just restart only one BGP instance and to trigger the graceful restart function? Thanks.

1 0