<p dir="ltr"><br>
23 июля 2015 г. 22:57 пользователь "Israel G. Lugo" <<a href="mailto:israel.lugo@tecnico.ulisboa.pt">israel.lugo@tecnico.ulisboa.pt</a>> написал:<br>
><br>
> Hello,<br>
><br>
> I am seeing periodic "Netlink: No such process" messages from bird6,<br>
> apparently related to prefixes learned from OSPF where there are two<br>
> equal cost paths. I've got ECMP on, and am installing routes to the<br>
> kernel via kernel protocol.<br>
><br>
> Operating system is GNU/Linux (Debian), kernel 3.16.<br>
><br>
> The problem happens when kernel protocol starts pruning table master.<br>
> Apparently, it's always deciding to update these multiple-path routes,<br>
> and has some failure while trying. Log excerpt:<br>
><br>
><br>
> bird6: kernel1: fec0:0:0:ffff::2/128: seen<br>
> bird6: kernel1: fec0:0:0:ffff::3/128: seen<br>
> bird6: kernel1: ::/0: seen<br>
> bird6: kernel1: Pruning table master<br>
> bird6: kernel1: 2001:xxx:yyy:z1::/64: updating<br>
> bird6: Netlink: No such process<br>
> bird6: kernel1: 2001:xxx:yyy:z2::/64: updating<br>
> bird6: Netlink: No such process<br>
> ...<br>
><br>
> This is what "show route" has to say about one of these routes:<br>
><br>
> bird> show route 2001:xxx:yyy:z1::/64 all<br>
> 2001:xxx:yyy:z1::/64 multipath [backbone 2015-07-21] * IA (150/20)<br>
> [193.aa.bb.137]<br>
> via fe80::21e:bff:fec1:8c4a on eth1 weight 1<br>
> via fe80::21e:bff:fec1:8c50 on eth1 weight 1<br>
> Type: OSPF-IA unicast univ<br>
> OSPF.metric1: 20<br>
> OSPF.metric2: 16777215<br>
> OSPF.tag: 0x00000000<br>
> OSPF.router_id: 193.aa.bb.137<br>
><br>
> "show route" periodically displays these routes with a '!' instead of<br>
> '*', indicating a synchronization error.<br>
><br>
> These "No such process" messages seem to occur every 40 seconds or so.<br>
><br>
> I am seeing this error both on BIRD 1.4.5 and BIRD 1.5.0.<br>
><br>
> Relevant topology: Four BIRD routers, A, B, C and D. All in area 0. A+B<br>
> are ABR for area 194, and C+D are ABR for area 165.<br>
><br>
> A and B announce IA prefixes z1, z2, z3, z4, z5, z6, z7 and z8 with<br>
> equal cost. C and D announce IA prefixes z9 and z10 with equal cost. A<br>
> and B have the error on prefixes from C and D, and vice-versa.<br>
><br>
> Routes seem to disappear from kernel periodically, and are reinstalled<br>
> again (monitoring with "ip -6 r").<br>
><br>
> Other routes, non-multipath from other routers, do not seem affected.<br>
><br>
> Relevant config, from routers C and D:<br>
><br>
> # common to all routers<br>
> protocol kernel {<br>
> learn;<br>
> persist;<br>
> scan time 20;<br>
> export all;<br>
> }<br>
><br>
> protocol ospf backbone {<br>
> tick 1;<br>
> ecmp yes;<br>
> area 0.0.0.0 {<br>
> stub no;<br>
> interface "eth1" { check link yes; };<br>
> };<br>
> area 0.0.165.0 {<br>
> stub yes;<br>
> summary yes;<br>
> interface "eth0.2000" {<br>
> type ptp;<br>
> check link yes;<br>
> };<br>
> interface "eth0.141", "eth0.165", "eth0.1411" {<br>
> stub;<br>
> check link yes;<br>
> };<br>
> networks {<br>
> 2001:xxx:yyy:z9::/64;<br>
> 2001:xxx:yyy:z10::/64;<br>
> };<br>
> };<br>
> }<br>
><br>
> Routers A and B are similar but different prefixes of course.<br>
><br>
> I do not see this problem with IPv4 bird (also OSPF, similar<br>
> configuration). Could this be some bug with kernel protocol and<br>
> multipath routes?<br>
><br>
> I am available for further explanations or more details (logs, configs).<br>
><br>
> Best regards,<br></p>
<p dir="ltr">Bird6 does not support ecmp in ipv6. It install and delete routes again on every sync. We hope that sometimes developers fix this issue.</p>
<p dir="ltr">><br>
> --<br>
> Israel G. Lugo<br>
> Núcleo de Redes e Comunicações<br>
> Direção de Serviços de Informática<br>
> Instituto Superior Técnico<br>
><br>
</p>