- Bird-users - Mailing Lists

BIRD memory usage
by Alexander Frolkin 01 Oct '15

01 Oct '15

Hello, We are running BIRD on a number of servers. It is configured with two BGP peers. We are seeing BIRD using over a gig of memory, and this seems excessive, especially given the number of routes. BIRD 1.5.0 ready. bird> show memory BIRD memory usage Routing tables: 1142 MB Route attributes: 48 kB ROA tables: 192 B Protocols: 45 kB Total: 1142 MB bird> show route count 656 of 656 routes for 333 networks The memory usage goes up slowly, starting from a few MB; this particular instance has been up for 3 days and 2 hours. Is there something we can do to reduce the memory usage? Or could this be a memory leak bug? The config is attached. Let me know if you need any more information. Thanks in advance! Alex

2 4

OSPF: directly attached summarized networks announced with metric 0 when link goes down
by Israel G. Lugo 01 Oct '15

01 Oct '15

Hello, I have a question regarding BIRD's OSPF and summarized networks. Not sure if I'm doing the right thing. I've got access routers running BIRD, configured as ABR between area 0 and their respective user-facing areas. Access networks are VLAN interfaces, e.g. eth0.161. Backbone connection is a separate physical interface (eth1). If the physical access interface (eth0) goes link down for some reason, BIRD changes the VLAN OSPF interfaces (eth0.161) to Loopback state. It stops announcing the directly connected prefixes on the VLAN interfaces, but it keeps announcing a /32 for the interface's IP, with a metric of 0. That in itself isn't the problem. It makes sense, as the IP belongs to the machine. The problem is, if I use the "network" option (to define summary LSAs), the /32 will be summarized to the entire prefix, making it be announced again. What's worse, it is now announced with a metric of 0, which means it will override any redundant routers I might have. Perhaps a simplified example will help explain: protocol ospf campus { tick 1; ecmp yes; area 0.0.0.0 { stub no; interface "eth1" { check link yes; }; # export loopback address into OSPF interface "lo" { stub; }; } area 0.0.161.0 { stub yes; summary yes; interface "eth0.161", "eth0.197" { stub; check link yes; }; networks { # eth0.161 A.B.C.0/24; # eth0.197 D.E.F.0/24; } } Access interfaces are eth0.161 and eth0.197, respectively with IPs A.B.C.253/24 and D.E.F.253/24. These networks are announced as A.B.C.0/24 and D.E.F.0/24, with a default metric of 10. Now, if I physically pull the cable from eth0, the result is that A.B.C.0/24 and D.E.F.0/24 start to be announced with metric 0, effectively sinking all traffic to those networks. [1] If I remove the "networks" option, no summarization is done, and things work: when I remove the cable from eth0, BIRD stops announcing the /24's and starts announcing A.B.C.253/32 and D.E.F.253/32 (with metric 0). I understand that literally, BIRD is doing what I asked it to do: summarize A.B.C.0/24. As long as it has some valid route inside that prefix (in this case the /32), it will announce the whole summarized network, with a metric equal to the largest cost (RFC 2328, 12.4.3). It would seem to me, though, that this case warrants special treatment. The /32 only exists because the interface transited to Loopback state when it lost the link. Perhaps I am looking at this the wrong way. I would like to summarize my networks, since I have several more complex cases where a shorter prefix is subnetted into several longer ones. However, I would also like to not announce dead routes from downed links. I am used to things working like this from other routers (Quagga, for example) [2], but perhaps there is a better way to do it in BIRD? Is there something I should be doing differently? Or could this perhaps be a bug, and not intended behavior? [1] As far as I can ascertain, what's happening is: eth0 goes link DOWN, eth0.161 and eth0.197 go LOWERLAYERDOWN, and BIRD puts them in Loopback state. As such, BIRD starts announcing their IPs as /32 with metric 0. Since it now has a /32 inside the summary network, it will start announcing the entire summary network, with the metric of 0. [2] In the case of Quagga for example, when the physical interface goes link down, all directly attached routes are removed. The /32 loopback addresses are not left behind. Hence, there is nothing to summarize, hence it will not announce the faulty prefix. Regards, -- Israel G. Lugo Núcleo de Redes e Comunicações Direcção de Serviços de Informática Instituto Superior Técnico

3 7

Support for RFC5838
by Ruben Herold 30 Sep '15

30 Sep '15

hi, I haven't found anything in the documentation. So I will ask the list, is there support for RFC5838 (Support of Address Families in OSPFv3) in bird? thx Ruben -- Ruben Herold ruben(a)insecure.pw

2 1

Missing BFD session with OSPFv3 and same link-local addresses.
by Christos Trochalakis 29 Sep '15

29 Sep '15

Hello all, Our issue is reproducible with 2 routers connected with 2 links. Those links are tagged vlans on the same physical interface (so both links share the same macaddress). bird> show ospf neighbors ospf1: Router ID Pri State DTime Interface Router IP B.B.B.B 1 full/other 00:08 vlan3509 fe80::d6ae:52ff:febf:74a6 B.B.B.B 1 full/dr 00:08 vlan106 fe80::d6ae:52ff:febf:74a6 bird> show bfd sessions bfd1: IP address Interface State Since Interval Timeout fe80::d6ae:52ff:febf:74a6 vlan3509 Up 2015-09-23 0.100 0.500 There is only one bfd session instead of two (one for each link). After digging a bit in the source code, it seems that the OSPF protocol calls `bfd_request_session()` which leads to a call to `bfd_add_request()`. `bfd_add_request()` tries to reuse an already established session with **the same destination address** before creating a new one. But B.B.B.B has the same link-local address on both interfaces (they share the same mac) so the second session is never created and the first one is returned. Perhaps a better approach would be to use the (address, iface) pair when trying to reuse a session. thank you, chris

2 1

Device protocol not recognizing interface address change
by Bernardo Figueiredo 29 Sep '15

29 Sep '15

Hello, When restarting some routers at the institution I work, sometimes bird doesn't start announcing any routes until we either restart bird or restart the device protocol using birdc. I think this happens due to the init scripts responsible for setting the network up finishing before the interfaces are fully configured(without the ip address set up), which allows for the init scripts dependent of the network init script, like bird, to start running. The problem is that bird detects that the interfaces are there, but doesn't recognize that the machine is configured with ip addresses(because the interfaces aren't fully configured yet) in the networks that it is announcing, so it doesn't announce the routes. When this happens: "bird> show ospf (...) Area networks: 10.3.0.0/16 Advertise 10.0.0.0/16 Advertise (...) " ( The problem is that the Area networks aren't active). After restarting the bird process, the output becomes correct(because now bird is starting with the interfaces fully configured): "bird> show ospf Area networks: 10.3.0.0/16 Advertise Active 10.0.0.0/16 Advertise Active " We have the device protocol to scan every 10 seconds in our bird.conf: "protocol device { scan time 10; }" I think that even though when bird first starts it shouldn't immediately start announcing(because it didn't see that the interfaces ip addresses), it should after 10 seconds when it scans again, because then the interfaces are already fully configured. For example, when radvd starts it complains that the interfaces don't have ip addresses configured, so it checks every second for the interfaces until they are fully configured, and the starts running correctly. Our current working workaround is make the networking init script wait for some time before exiting, meaning that when the init dependents start executing, the interfaces are already fully configured. This might work for the machine startup, but it doesn't solve the problem when we add a new ip address in a new network to an already existing interface and make bird start announcing that network without restarting bird (or the interface). I haven't checked that the same behaviour described above happens in this situation, but it should. If you want me to try it I can. If I haven't explained some parts clearly just ask and I'll try to explain it better. Also, if you need any configuration details just ask. This problems happens in both bird "1.4.5-1~bpo70+1" and "1.5.0-4" distributed by Debian. So, my question is, shouldn't the Device protocol scan update see that the ip addresses of an interface have changed and if so, start announcing the route. Thank you, Bernardo Figueiredo

3 11

Fwd: bird-bug (cpu load 100%)
by Мамонов Павел 24 Sep '15

24 Sep '15

Good afternoon, Faced with the problem of initialization, if the option is: "protocol device {scan time of 10}" the server hangs with 100% CPU-loading, if you remove this option the server boots normally. You could not tell in what could be the problem. Server - 3.18.21-Gentoo, ixgbe - 4.1.2 (10G SFP) -- С уважением, Мамонов Павел, сетевой администратор ООО "Астрателком" тел. (099)2419158 г.Одесса, ул. Бугаевская, 21/23 e-mail: pavel.ma(a)astra.od.ua

1 0

RIP v2 'router id' significance
by Mark Trumpold 22 Sep '15

22 Sep '15

Hello, In a rip domain with multiple routers in a hub/star configuration, is is possible/advisable not to specify a ‘router id’ in the router configs? Thanks in advance, Mark T.

2 1

Bird BGP issue over stringswan
by Ajay Raut ( Tech Ops ) 21 Sep '15

21 Sep '15

Hi, My setup is BGP over a ipsec tunnel. Most of the time it works fine but in between on config soft reload, it goes into a error state. So its intermittent issue where sometime it works. Debug logs show bad BGP as error but same configuration starts working intermittently, otherside of the bird is Juniper SRX endpoint Log error message are: 015-09-21 16:32:22 <TRACE> evpn: Connecting to 10.11.201.1 from local address 10.11.201.13 2015-09-21 16:32:24 <TRACE> evpn: Connected 2015-09-21 16:32:24 <TRACE> evpn: Sending OPEN(ver=4,as=65529,hold=240,id=0a0bc90b) 2015-09-21 16:32:25 <TRACE> evpn: Got OPEN(as=0,hold=90,id=0a54c804) 2015-09-21 16:32:25 <RMT> evpn: Error: Bad peer AS: 0 2015-09-21 16:32:25 <TRACE> evpn: State changed to stop 2015-09-21 16:32:25 <TRACE> evpn: Sending NOTIFICATION(code=2.2) 2015-09-21 16:32:25 <TRACE> evpn: Down 2015-09-21 16:32:25 <TRACE> evpn: State changed to down any help on this will be very useful. Thanks, Ajay

1 0

OSPF export/import default route filter problem
by LU 20 Sep '15

20 Sep '15

Hi, I have 4 routers connected like this. (10.4.4.1/24) (10.4.4.4.2/24) R1 R2 \ / \ / SWITCH / \ / \ R3 R4 (10.4.4.3/24) (10.4.4.4/24) I export a default route (0.0.0.0/0) from R1 and R2 with an ospf_metric1 = 100.. At R3, in an import filter I set the ospf_metric1 of R1's default route to 10. At R4, in an import filter I set the ospf_metric1 of R2's default route to 10. What I expect is that R3's routing table is going to contain 0.0.0.0/0 via R1. and likewise R4's routing table is going to contain 0.0.0.0/0 via R2. Unfortunately it doesn't happen and both R3's and R4's routing table has 0.0.0.0/0 via R2. Here's a config for router R1 (R2 is identical except for router id). router id 10.4.4.1 protocol kernel { export all; } protocol direct { } protocol device { scan time 10; } protocol static { route 0.0.0.0/0 via 123.123.123.1; #vtnet0 } protocol ospf { tick 2; rfc1583compat yes; export filter { if net = 0.0.0.0/0 then { ospf_metric1 = 100; accept; } reject; } area 0.0.0.0 { interface "vtnet1" { dead count 4; type broadcast; } } } Here's a config for router R3 (R4 is identical except for router id and if statement's ospf_router_id) router id 10.4.4.3 protocol kernel { export all; } protocol direct { } protocol device { scan time 10; } protocol static { } protocol ospf { tick 2; rfc1583compat yes; import filter { if ospf_router_id = 10.4.4.1 then { if net = 0.0.0.0/0 then { ospf_metric1 = 10; } } accept; } area 0.0.0.0 { interface "vtnet0" { dead count 4; type broadcast; } interface "ng*" { stub yes; } } } When I place a print statement in the import filter at the very top: import filter { print "router_id: ", ospf_router_id, " net: ", net; if ospf_router_id = 10.4.4.1 then { if net = 0.0.0.0/0 then { ospf_metric1 = 10; } } For R3 I see: router_id: 10.4.4.3 net: 10.4.4.0/24 router_id: 10.4.4.2 net: 0.0.0.0/0 router_id: 10.4.4.2 net: 10.4.4.0/24 For R4 I see: router_id: 10.4.4.4 net: 10.4.4.0/24 router_id: 10.4.4.2 net: 0.0.0.0/0 router_id: 10.4.4.2 net: 10.4.4.0/24 So router R1 (10.4.4.1) is not seen by the import filters, as such R2 is selected as the default route for both R3 and R4. How could i fix it? Thank you..

2 2

pmacct and BIRD on the same box
by Charles van Niman 20 Sep '15

20 Sep '15

Hello, I am looking to feed BGP data into pmacct on a linux router I have running BIRD. Right now, BIRD runs iBGP to five other boxes [quagga/BIRD mix], on one IP [172.22.150.126], assigned to the loopback interface, and a handful of eBGP peers on another IP [172.22.150.127]. I see two practical approaches to this, but I am all ears for more! My original plan was to just put pmacct's BGP peer IP on a separate loopback IP [172.22.150.119], and peer between BIRD 172.22.150.126:179 and pmacct 172.22.150.119:179, however since I am using two different listening IPs in BIRD, it binds to 0.0.0.0. I attempted to use the global listen bgp option, but it seems I am forced to pick between eBGP, iBGP, running multiple BIRD instances, or renumbering sessions, which I want to avoid :( Another way around this seemed to be to use a different TCP port for pmacct to bind to, however BIRD doesn't seem to support per-neighbor port settings. Maybe I have just missed this feature, if so hooray! Please let me know any thoughts or ideas you have on this, I am open to anything, including how I may have worded this better. Regards, Charles

2 2