Hello,
Yes it is iBGP, by default next hop for ipv6 route is set to link-local ipv6 
address, in my setup I need global ipv6 address as next-hop. My config works 
but not for every peer under the same bgp template, I have to reload ipv6 bgp 
protocol to fix it.

Dariusz

On 7/22/20, 1:12 PM, "Alexander Zubkov" <[email protected]> wrote:

    Hi,
    
    Looks like you have the same AS on your peers. So they have iBGP
    sessions. And those keep the next hop attribute by default.
    
    On Wed, Jul 22, 2020 at 12:55 PM Mazur, Dariusz <[email protected]> wrote:
    >
    > Hello Bird Users,
    >
    > Hope you will be able to help me. I have 8  bgp peers which use  the same 
bgp template with below import filter:
    >
    >
    >
    > filter IMPORT__SPINE__TO__LEAF
    >
    > {
    >
    >     bgp_community.add(LOOP_DETECTION);
    >
    >     gw = bgp_next_hop;  -> I use it to set ipv6 bgp next hop to global 
ipv6 address  (not link local)
    >
    >     if (net = 0.0.0.0/0 || net = ::/0) then
    >
    >     {
    >
    >         reject;
    >
    >     }
    >
    >    …
    >
    >     else
    >
    >     {
    >
    >         reject;
    >
    >     }
    >
    > }
    >
    >
    >
    > 1.First Peer  -> this look good
    >
    >
    >
    > r14.spine101.fra02.fab> bird show protocols all 
2a02:26f0:3700:1101::1__r01.leaf101
    >
    > BIRD 2.0.7 ready.
    >
    > Name       Proto      Table      State  Since         Info
    >
    > 2a02:26f0:3700:1101::1__r01.leaf101 BGP        ---        up     
09:58:55.848  Established
    >
    > …
    >
    > Channel ipv6
    >
    >     State:          UP
    >
    >     Table:          master6
    >
    >     Preference:     100
    >
    >     Input filter:   IMPORT__SPINE__TO__LEAF
    >
    >     Output filter:  EXPORT__SPINE__TO__LEAF
    >
    >     Routes:         27 imported, 4 exported, 25 preferred   -> I expect 
to import 27 routes
    >
    >     Route change stats:     received   rejected   filtered    ignored   
accepted
    >
    >       Import updates:             27          0          0          0     
    27
    >
    >       Import withdraws:            0          0        ---          0     
     0
    >
    >       Export updates:          15846         50      15791        ---     
     5
    >
    >       Export withdraws:           11        ---        ---        ---     
     0
    >
    >     BGP Next hop:   2a02:26f0:3700:1101::
    >
    >
    >
    > 2.Second Peer -> this does not look good:
    >
    >
    >
    > r14.spine101.fra02.fab> bird show protocols all 
2a02:26f0:3700:1102::1__r02.leaf101
    >
    > BIRD 2.0.7 ready.
    >
    > Name       Proto      Table      State  Since         Info
    >
    > 2a02:26f0:3700:1102::1__r02.leaf101 BGP        ---        up     
09:58:56.676  Established
    >
    >
    >
    >   Channel ipv6
    >
    >     State:          UP
    >
    >     Table:          master6
    >
    >     Preference:     100
    >
    >     Input filter:   IMPORT__SPINE__TO__LEAF
    >
    >     Output filter:  EXPORT__SPINE__TO__LEAF
    >
    >     Routes:         4 imported, 4 exported, 0 preferred  -> bird imports  
only 4 routes
    >
    >     Route change stats:     received   rejected   filtered    ignored   
accepted
    >
    >       Import updates:             27          0                           
  23          0          4       -> bird reject 23 routes
    >
    >
    >
    > After I reload this protocol all looks good and bird import 27 routes
    >
    >
    >
    > In logs I see  “Invalid gw address”,  therefore I suspect issue with “gw 
= bgp_next_hop”
    >
    >
    >
    > 2020-07-22 09:58:56.801 <ERR> filters, line 169: Invalid gw address
    >
    > 2020-07-22 09:58:56.801 <ERR> filters, line 169: Invalid gw address
    >
    > 2020-07-22 09:58:56.801 <ERR> filters, line 169: Invalid gw address
    >
    > 2020-07-22 09:58:56.801 <ERR> filters, line 169: Invalid gw address
    >
    > 2020-07-22 09:58:56.801 <ERR> filters, line 169: Invalid gw address
    >
    >
    >
    > I am able to reproduce issue on more routers, please find below out from 
another router:
    >
    > r16.spine101.fra02.fab> show bgp summary  ->  it parse “show protocols 
all”
    >
    >
    >
    > IPv6 Peers: 10 Down: 0
    >
    > Peer                          AS      State       (Im/Filt/Exp/Pref)      
  Up/Down         Protocol
    >
    > 2a02:26f0:3700:211::          20940   Established (8480/0/35/8480)        
  00:36:46        2a02:26f0:3700:211::__r01.border101
    >
    > 2a02:26f0:3700:311::          20940   Established (7131/0/34/7)           
  00:36:48        2a02:26f0:3700:311::__r02.border101
    >
    > 2a02:26f0:3700:1301::1        20940   Established (28/0/4/26)             
  00:36:51        2a02:26f0:3700:1301::1__r01.leaf101
    >
    > 2a02:26f0:3700:1302::1        20940   Established (4/0/4/0)               
  00:36:49        2a02:26f0:3700:1302::1__r02.leaf101  -> 4 routes are imported 
but should be 28
    >
    > 2a02:26f0:3700:1303::1        20940   Established (28/0/4/1)              
  00:36:53        2a02:26f0:3700:1303::1__r03.leaf101
    >
    > 2a02:26f0:3700:1304::1        20940   Established (4/0/4/0)               
  00:36:50        2a02:26f0:3700:1304::1__r04.leaf101
    >
    > 2a02:26f0:3700:1305::1        20940   Established (28/0/4/1)              
  00:36:47        2a02:26f0:3700:1305::1__r05.leaf101
    >
    > 2a02:26f0:3700:1306::1        20940   Established (28/0/4/1)              
  00:36:46        2a02:26f0:3700:1306::1__r06.leaf101
    >
    > 2a02:26f0:3700:1307::1        20940   Established (28/0/4/1)              
  00:36:47        2a02:26f0:3700:1307::1__r07.leaf101
    >
    > 2a02:26f0:3700:1308::1        20940   Established (28/0/4/1)              
  00:36:46        2a02:26f0:3700:1308::1__r08.leaf101
    >
    >
    >
    > After protocol reload all looks good.
    >
    >
    >
    > Did you face similar problem?
    >
    >
    >
    > Regards,
    >
    > Dariusz
    >
    >
    >
    >
    >
    >
    >
    >
    >
    >
    >
    >
    >
    >
    


Reply via email to