- Bird-users - Mailing Lists

date & time change
by Savino Rapillo 07 Feb '05

07 Feb '05

Hi All I'm new in this list, and I'm very sorry to bother you with this question anyway if someone has already met it I'll be glad to learn something new. bird seams to be really disturbed by a change in date/time I have a net of linux box linked via WiFi and, of course, their routing table are handled with bird (ospf). We are using version 1.0.7 (but with the 1.0.9 that problem is still there). Well, when there is a change in date/time (set via ntpd or manual), bird seems to lock in a state insane. What I can see is that the routing tables are removed and the only thing that can restart the job seems to be a kill of bird. I have some idea of what can appen to bird timers when someone plays with "date" but I wander if someone can suggest some smart idea (or can confirm that there is nothing to do) to work around the problem. Thank you Savino

3 2

Route change trouble with bird
by michael_moore＠csnw.com 21 Jan '05

21 Jan '05

Greetings: I am beginning to use bird 1.0.9 under Linux 2.2 on a small IPv4-only network and uncovered a difficulty with using the "ip route change" command. The linux router in question is configured with "protocol kernel" having "preference 180" so that local "alien" routes will supercede equivalent RIP routes (so that backup routes managed by the linux router will be taken instead of routes available on other routers). I noticed that if I try to use the "ip route change ..." command to modify an "alien" route, the route is instead removed and replaced by the RIP route. I can use the same command a second time to replace the RIP route with the "alien" route and it works properly. Also, if I first delete the original "alien" route, wait a second for the RIP route to take its place, then issue the "ip route change..." command, the change works properly. For example: - The route (via "ip route list") is: 206.163.40.96/27 via 65.100.46.254 dev ipsec0 - I execute: ip route change 206.163.40.96/27 via 65.100.46.254 dev ipsec0 src 206.163.40.44 - The following items appear in the log: 20-01-2005 16:14:55 <TRACE> kernel1: 206.163.40.96/27: [alien async] removed 20-01-2005 16:14:55 <TRACE> rip1 < forced accept by protocol 206.163.40.96/27 via 206.163.40.33 on ipsec1 20-01-2005 16:14:55 <TRACE> rip1 < replaced 206.163.40.96/27 via 206.163.40.33 on ipsec1 20-01-2005 16:14:55 <TRACE> kernel1 < replaced 206.163.40.96/27 via 206.163.40.33 on ipsec1 20-01-2005 16:14:55 <TRACE> kernel1 > removed [replaced] 206.163.40.96/27 via 65.100.46.254 on ipsec0 20-01-2005 16:14:55 <TRACE> kernel1: 206.163.40.96/27: [alien async] created 20-01-2005 16:14:55 <TRACE> kernel1 > added [best] 206.163.40.96/27 via 65.100.46.254 on ipsec0 20-01-2005 16:14:55 <TRACE> rip1 < replaced 206.163.40.96/27 via 65.100.46.254 on ipsec0 20-01-2005 16:14:55 <TRACE> kernel1 < replaced 206.163.40.96/27 via 65.100.46.254 on ipsec0 20-01-2005 16:14:55 <TRACE> kernel1: 206.163.40.96/27: [alien async] removed 20-01-2005 16:14:55 <TRACE> rip1 < forced accept by protocol 206.163.40.96/27 via 206.163.40.33 on ipsec1 20-01-2005 16:14:55 <TRACE> rip1 < replaced 206.163.40.96/27 via 206.163.40.33 on ipsec1 20-01-2005 16:14:55 <TRACE> kernel1 < replaced 206.163.40.96/27 via 206.163.40.33 on ipsec1 20-01-2005 16:14:55 <TRACE> kernel1 > removed [replaced] 206.163.40.96/27 via 65.100.46.254 on ipsec0 - The route (via "ip route list") now shows a totally different (RIP) route: 206.163.40.96/27 via 206.163.40.33 dev eth1 proto bird Looking at the code, I couldn't spot the trouble spot. I am left to imagine that the timing of the "ip route change" command conflicts with the timing of bird adding its own RIP route. I just don't know how it works in the Netlink interface. As indicated above, preceding the route change command with a delete and a delay results in the change working properly: ip route del 206.163.40.96/27 sleep 1 ip route change 206.163.40.96/27 via 65.100.46.254 dev ipsec0 src 206.163.40.44 I noticed lots of low-level debugging output in the netlink code, but I couldn't figure out how to output it. When I configure with "--enable-debug", I don't see any additional output in the debug file or or stderr (with "-d") or in the "debug output file" (with "-D bugout"), but I do have to specify the locations of the bird.conf and bird.ctl files. If you have any hints for debugging, I'm eager for those as well. Thanks, -Michael

1 0

Troubles with aliased interfaces
by Nicolas Olivier 12 Dec '04

12 Dec '04

Hello, I'm getting trouble to setup bird with aliases. When a bird is bound on an IP assigned to an alias, it will answer with the IP of the interface, not the alias one, which is a bit ennoying. If someone got an easy way to do such setup, it will be greatly appreciated. Nicolas Olivier.

3 2

Propagating my IPs via BGP
by Michal Kurka 12 Dec '04

12 Dec '04

Hi. I'm assuming that this list is in English only (no czech). I use Bird for BGP-routing. I need to propagate to my ISP a new IP-segment (e.g. 100.101.102.0/24) newly assigned by RIPE. My config-file contains: ------------------------------------------------------------------ log "/var/log/bird.log" all; protocol kernel { persist; scan time 20; learn off; export all; } protocol device { scan time 10; } protocol bgp { local as $MY_AS; neighbor $ISP_PEER as $ISP_AS; export all; } ------------------------------------------------------------------ I assigned IP 100.101.102.1/24 to an eth-device (direction to LAN) and I expected that Bird make scan a device and take an IP to BGP. But my ISP says that nothing received. Where have I a mistake? Thanks for answer. -- Michal Kurka - Mysak spravce serveru 'kapsa.club.cz'

2 1

FAT bird again
by Zharovtsev, Anton 12 Dec '04

12 Dec '04

Hi all. Time ago I've asked the question about growing birds memory. Few days ago I've found dirds debug output :-). Here is all dumps in attached archieve (tar.gz) In short: just right after start bird has less then 2M of memory in my system, Next day memory becomes 37M and after two days of uptime - 64M. At the another machine bird does not growing ... The differens is only in firewall rules - I don't accept packets from untrusted RIP neighbors.

4 5

ospf ecmp implementation - what to do?
by Birger Schmidt 21 Oct '04

21 Oct '04

hi all, i want to have multipath support in bird. but as in 1.0.9 it isn't. so i read the source an find in proto/ospf/rt.c a comment that say: /* ... we don't support multiple next hops. */ /* I'll start as soon as nest will */ so i looked at the core of bird called nest. but unfortunately it is not clear to me where to start. i think support has to go to nest/rt-table.c but if someone would give me some advise what has to be done exactly that would be very nice. what do you think how complicated is it to implement multipath functionality in bird? thanks birger

1 0

RE: FAT bird again
by Zharovtsev, Anton 12 Oct '04

12 Oct '04

They all have the same work style: they all begin to grow when I turn off my firewall that accept only trusted host. Befor I've tried 1.0.6, 1.0.8 and finaly 1.0.9. Logs are from bird 1.0.9. Bird does not grow from my test BGP session with zebra, with bird filters or without it. When I turn on firewall this concrete specimen of bird does not grow. There are too mutch crazy and enemy RIPs in my ethernet segment, they are from different daemons, from Windows machines, RIPv1, RIP from no-neighbors (other ip subnets) and so on ... Can it be the reason ? -----Original Message----- From: Ondrej Filip [mailto:feela@network.cz] Sent: Tuesday, October 12, 2004 3:18 AM To: Zharovtsev, Anton Cc: bird-users(a)bird.network.cz Subject: Re: FAT bird again > At the another machine bird does not growing ... The differens is only > in firewall rules - I don't accept packets from untrusted RIP > neighbors. > Are all birds version 1.0.9? F. >

1 0

Bgp filtering
by Zharovtsev, Anton 07 Oct '04

07 Oct '04

Hi all. Tell me please how should I write filtern of bgpmask in the birdc ? I want to write something like: bird> show route table testtable where bgp_path ~ /? 65260 ?/ But when I typing the "?" birdc gives me the help :-). Please tell me how to see full bgp_path or give me some examples with it inputs and outputs. What should I type in birdc to see all routes in table testtable which have 65260 in their bgp_path for example ? I suspect that is possible :-). P.S. I don't anderstand BGP very good :-(.

1 0

Re: bird vs zebra
by Ramon Martinez Palomares 16 Sep '04

16 Sep '04

still having any problems with zebra and OSPF. is anybody using bird to intercommunicate with zebra in a pointopoint link? bird is connected to the zebras but i cant see any route coming from bird to zebra. (sorry if this is a FAQ but i am new to the list and i havent seen an archive) El 15/09/2004, a las 21:07, Ramon Martinez Palomares escribió: > > Hello all! > > while trying to comunicate zebra and bird in a tunnel (pointopoint) > conection, all the time bird keeps saying that: > > Sep 15 21:06:20 bell bird: Bad OSPF hello packet from 172.16.72.157 > received: bad netma > sk 0.0.0.0. > > > after looking at hello.c i have commented the return after the log > message. and bird is working really fine. > > What this error really means? is it a bird or a zebra issue? > > > -ramon- > >

2 2

Re: bird vs zebra
by Ondrej Filip 15 Sep '04

15 Sep '04

On Wed, 15 Sep 2004, Ramon Martinez Palomares wrote: > > Hello all! Hello! > > while trying to comunicate zebra and bird in a tunnel (pointopoint) > conection, all the time bird keeps saying that: > > Sep 15 21:06:20 bell bird: Bad OSPF hello packet from 172.16.72.157 received: > bad netmask 0.0.0.0. > > > after looking at hello.c i have commented the return after the log message. > and bird is working really fine. > > What this error really means? is it a bird or a zebra issue? It is a bird's bug. It will be fixed in next release. Thank you! Ondrej > > > -ramon- > >

1 0