- Bird-users - Mailing Lists

Route Servers and Secondary Routes (bird 1.3.10 and debian kernel 3.2-amd64)
by Daniel Gomez 25 Nov '13

25 Nov '13

Hi Everyone, I have used both examples from Ondrej "Simple Route Server" and "Router server with community based filtering and single RIB", from the gitlabs.labs.nic.cz, to configure 2 Route Servers that need to tranfer to each other via iBGP all the routes they have learned. I thought that the most important parameters, when coupling both Router Servers, where the "table master sorted" and the "secondary" parameter for the iBGP configuration, but my routing table is just showing the Best Route and not all of them. I am pushing to one of the servers via iBGP and "bgpdump/bgpsimple" around 775775 prefixes (www.ripe.net/cata-tools/stats/ris/ris-raw-data) I do not know why but the route server is receiving the 775775 prefixes, filtering them and just showing the preferred ones. I have tried with "show route all filtered", but the command does not show anything. R65500x1 BGP master up 14:11 Established Description: 192.168.0.1 Preference: 100 Input filter: ACCEPT Output filter: ACCEPT Routes: 47423 imported, 0 exported, 47423 preferred <<<---- Route change stats: received rejected filtered ignored accepted Import updates: 775775 0 0 17521 758254 <<<---- Import withdraws: 6 0 --- 2 4 Export updates: 758255 758254 2 --- 0 Export withdraws: 4 --- --- --- 0 BGP state: Established Neighbor address: 192.168.0.1 Neighbor AS: 65500 Neighbor ID: 192.168.0.1 Neighbor caps: Session: internal Source address: 192.168.0.2 Hold timer: 1905/2400 Keepalive timer: 339/800 // CONFIGURATION log "/var/log/bird.log" all; log syslog all; router id 192.168.0.2; define myas = 65500; #### # Protocol template template bgp PEERS { local as myas; source address 192.168.0.2; secondary; import all; export all; route limit 10000; } table master sorted; #### # Configuration of BGP peer follows ### AS65500 - Member1 protocol bgp R65500x1 from PEERS { description "192.168.0.1"; gateway direct; neighbor 192.168.0.1 as 65500; } // LOGS Nov 25 17:06:12 bird: R65500x1 > added [best] 1.0.4.0/24 via 192.168.0.1 on eth2 Nov 25 17:06:12 bird: kernel1 < added 1.0.4.0/24 via 192.168.0.1 on eth2 Nov 25 17:06:12 bird: R65500x1 < rejected by protocol 1.0.4.0/24 via 192.168.0.1 on eth2 Nov 25 17:06:12 bird: R65500x1 > added [best] 1.0.4.0/24 via 192.168.0.1 on eth2 Nov 25 17:06:12 bird: kernel1 < replaced 1.0.4.0/24 via 192.168.0.1 on eth2 Nov 25 17:06:12 bird: R65500x1 < rejected by protocol 1.0.4.0/24 via 192.168.0.1 on eth2 Nov 25 17:06:12 bird: R65500x1 > added [best] 1.0.4.0/24 via 192.168.0.1 on eth2 Nov 25 17:06:12 bird: kernel1 < replaced 1.0.4.0/24 via 192.168.0.1 on eth2 Nov 25 17:06:12 bird: R65500x1 < rejected by protocol 1.0.4.0/24 via 192.168.0.1 on eth2 Nov 25 17:06:12 bird: R65500x1 > added [best] 1.0.4.0/24 via 192.168.0.1 on eth2 Any idea? Thanks in advance, Daniel Gomez

2 2

BIRD on FreeBSD 9.2 (amd64) - OSPF is not putting routes in the routing table
by Claude Marinier 21 Nov '13

21 Nov '13

Hi, I am running BIRD 1.3.11 on FreeBSD 9.2 (amd64). It establishes full adjacency with its three neighbours (Cisco routers) and its routing table is correct. BIRD knows about the routers behind its neighbours. The problem is that the kernel does not know any of the learned routes, i.e. 'netstat -rnf inet' shows only routes for the interfaces. I suspect that I am missing something basic but have not found it yet after many hours of searching. Could someone point me in the right direction? Thank you. The configuration file follows. -- C Marinier # Configure logging log syslog { debug, trace, info, remote, warning, error, auth, fatal, bug }; # Override router ID router id 131.140.113.14; protocol kernel { persist; # leave routes in kernel table on exit import all; # import routes into kernel routing table scan time 10; # look for new interfaces every 10 seconds } protocol device { scan time 10; # look for changes every 10 seconds } # make sure routes for directly connected sub-networks # are not disturbed (BSD's are susceptible to this) protocol direct { interface "bce0", "bce1", "em0"; } protocol ospf WANemu { area 0.0.0.0 { stub no; networks { x.x.113.8/29; x.x.113.184/29; x.x.113.104/29; }; interface "bce0" { authentication cryptographic; password "xxxxxx"; }; interface "bce1" { authentication cryptographic; password "xxxxxx"; }; interface "em0" { authentication cryptographic; password "xxxxxx"; }; }; }

3 3

I failed to compile the BFD branch
by 三ツ木　賢一 21 Nov '13

21 Nov '13

Hello all. I have compiled a BFD Branch. It is following hash. Commit 1ec522538fb81a56b068c087d0a842faf7aa7869 However, it was not possible and it is said that there is no "nest/bfd.h" Is there a solution? .

2 1

Mail function
by Anibe Onuche 20 Nov '13

20 Nov '13

Hello I am suggesting that BIRD should have a mail function that alerts the administrator of critical errors in real time especially from it's syslog severs. Bird runs on a Linux platform and can be incorporated with some mailer daemon If a BGP section goes down.. I want Bird to alert me, If there is a change on my configuration.. I should be alerted etc. Would be grateful -- Anibe Onuche

5 6

AS numbers higher than 2000000000
by Milan Kocian 18 Nov '13

18 Nov '13

hi, I tried set up private 4b AS number (4200000000-4294967294). And I've discovered that my filters using int datatype doesn't work with these numbers. The reason is 'int' size (from -2000000000 to +2000000000). Does exist some other type for it (I didn't find anything in doc.) ? Best regards, -- Milan Kocian

3 3

More IPSEC routes for OSPF
by Iain Buchanan 18 Nov '13

18 Nov '13

Hi, I've a question about getting OpenSWAN and Bird routing working together. Hopefully someone here can point me in the right direction! There was an email from Michael Ludvig on the 8th July 2013 about exporting IPSEC routes for OSPF usage. The conclusion seemed to be that Bird can’t read the routes set up by OpenSWAN, and the only solution is to have a script running that exports the routes for Bird’s use into a separate table. I’m in pretty much the same position. I’ve tried Ondrej Zajicek’s suggestion of using transport mode IPSEC links, but this doesn’t seem to create visible routes (I’m using the netkey stack, which may be the issue). At the moment I’ve got GRE tunnels working on top of the IPSEC links, and if I enable debugging mode I can see instances of Bird communicating with one another over them (but not sending any of the OpenSWAN link information). I’d like to try and create routes that Bird can read, but I’m not certain how to extract the routing information nicely. (Part of the problem may be that the IPSEC links go through VPNs.) If I do an “ip xfrm policy” the output only refers to the gateways (i.e. where the NAT takes place), not the machines that I want to gain access to. I think what I need is a route saying “for remote_network/mask use the default route on this machine”. This could be written to a dedicated routing table which bird would read and send on. I’m guessing packets would then turn up and be processed by the xfrm rules and head on through the tunnel. It looks like the way to do this is to have a little service running that will regularly: (1) parse output from “ip route” to determine the default route (2) parse output from “ip address” to determine the set of local IP addresses (3) parse output from “ipsec auto status” looking for any of the “network diagram” lines that show the connectivity where one end or the other goes through a local IP address (4) update the routing table with calls to "ip route" (5) somehow prod Bird so that it reads the table I hope I’ve horribly over-complicated things and there’s an easier way to do this… does this sound like the way to go? Iain

6 11

Re: simple filters question
by Sergey Popovich 18 Nov '13

18 Nov '13

В письме от 9 ноября 2013 07:42:06 Вы написали: > Hi, > > I applied this code manually: > > + if ((v1.type == T_IP) && (v2.type == T_PREFIX_SET)) { > + v1.val.px.len = MAX_PREFIX_LENGTH; > + return trie_match_fprefix(v2.val.ti, &v1.val.px); > + } > > To filter.c > > However maybe the patch failed because I don't have the following: > > if (v2.type != T_SET) > return CMP_ERROR; > > > Maybe the patch that was sent was within a different branch than the main > 1.3.11 that's available for download? > Oh sorry, John, I patch really was made on top of current upstream tree, not 1.3.11. In attachment udpated patch version applied on top of 1.3.11 stable tree. Sorry for that. And yes, really according to documentation we could match only on set with elements of same tyme. On the other hand, an ip type may be easily represented as prefix with length of max AFI prefix length (32 - IPv4, 128 - IPv6, so I see no problem using this conversion of ip to prefix type. Moreover on IPv4 BIRD build there is implicit conversion of ip type to quad as on IPv4 they basically represent value with same format. Thanks for your note. -- SP5474-RIPE Sergey Popovich

3 2

Route table max up to 32K
by Bryan Duff 18 Nov '13

18 Nov '13

Any reason why the max route table is limited to 256 and not the Linux limit of 32767? Also any thoughts of adding dbus support or something similar for direct command injection (instead of rewriting the config and HUP'ing)? Thanks. -Bryan

4 3

simple filters question
by John Jensen 08 Nov '13

08 Nov '13

Hi, We're about to migrate our exchange to a /23 and we're likely going to do it over time. I have a simple filter in place to check and make sure that the NEXT_HOP is within our exchange's subnet. Originally it looked like this: function check_ixp_next_hop(ip nexthop) prefix ixpnet; { ixpnet = x.x.x.x/24; if ! (nexthop ~ ixpnet) then return false; return true; } And then within the inbound filter for each peer's ASN: if ! (check_ixp_next_hop(bgp_next_hop)) then reject; Since we're going to be caring about two distinct prefixes while we perform the migration, I changed the filter to look like this: function check_ixp_next_hop(ip nexthop) prefix set ixpnet; { ixpnet = [ x.x.x.x/24, y.y.y.y/23 ]; if ! (nexthop ~ ixpnet) then return false; return true; } However I'm getting the following errors in the log: 07-11-2013 12:33:00 <ERR> filters, line 61: ~ applied on unknown type pair 07-11-2013 12:33:00 <ERR> filters, line 61: ~ applied on unknown type pair 07-11-2013 12:33:04 <ERR> filters, line 61: ~ applied on unknown type pair Am I missing something? Thanks, -JJ

2 8

How bird configure multiple instances on ospf?
by 曾小小 06 Nov '13

06 Nov '13

I am now in use BIRD in the ospf routing protocol, such as cisco router I want to use multiple processes on the ospf. This how to achieve it ? Such as cisco router multi-process examples: r1 (config) # router ospf 100 r1 (config-router) # network 192.168.3.0 0.0.0.255 area 0 r1 (config-router) # network 10.0.5.0 0.0.0.255 area 0 r1 (config-router) # exit r1 (config) # router ospf 101 r1 (config-router) # network 10.0.7.0 0.0.0.255 area 1 My configuration is as follows , but I do not know if this configuration is not multi-process approach. filter export_OSPF { if (source = RTS_STATIC && net = 0.0.0.0 / 0) then { print "net accepted:", net; ospf_metric2 = 20; accept; } reject; } # # protocol ospf ospf_1 { table tab_ospf_1; router id x.x.x.x debug all; export filter export_OSPF; import all; preference 200; area 0.0.0.0 { interface "gre0" { hello 10; retransmit 5; cost 10; transmit delay 1; dead count 4; wait 40; type pointopoint; # type pointomultipoint; # type broadcast; priority 0; authentication cryptographic; password "abc"; } ; } ; } # # protocol ospf ospf_2 { table tab_ospf_2; router id x.x.x.x; # Debug all; export filter export_OSPF; import all; preference 200; area 0.0.0.0 { interface "gre1" { hello 10; retransmit 5; cost 10; transmit delay 1; dead count 4; wait 40; type pointopoint; # type broadcast; # type pointomultipoint; priority 0; authentication cryptographic; password "abc"; } ; } ; } Thank you..

1 0