IPv6 multipath static routes with different weights and BFD supervision
Hi BIRD experts, I face some problems on static routes on IPv6. I try to configure multipath static routes that are supervised with BFD. I use the following conf: route <DEST_IPv6> multipath via <NEXT_HOP1_IPv6> weight 10 bfd via <NEXT_HOP2_IPv6> weight 20 bfd; This does not seem to have the expected behavior since I notice that the static routes are configured with the same priority rather than having the less weighted one with higher priority than the more weighted. When I try to check the routes passed to the kernel I notice this: <DEST_IPv6> via <NEXT_HOP1_IPv6> dev eth3 proto bird metric 1024 <DEST_IPv6> via <NEXT_HOP2_IPv6> dev eth5 proto bird metric 1024 The same configuration for IPv4 works as expected... could you, please, give me some initial orientation regarding the issue? The BIRD version is BIRD 1.6.3 on Debian GNU/Linux 8 (jessie)
On Mon, Jan 22, 2018 at 03:39:06PM +0000, ?????? ?????????? wrote:
Hi BIRD experts, I face some problems on static routes on IPv6. I try to configure multipath static routes that are supervised with BFD. I use the following conf: route <DEST_IPv6> multipath via <NEXT_HOP1_IPv6> weight 10 bfd via <NEXT_HOP2_IPv6> weight 20 bfd; This does not seem to have the expected behavior since I notice that the static routes are configured with the same priority rather than having the less weighted one with higher priority than the more weighted. When I try to check the routes passed to the kernel I notice this: <DEST_IPv6> via <NEXT_HOP1_IPv6> dev eth3 proto bird metric 1024 <DEST_IPv6> via <NEXT_HOP2_IPv6> dev eth5 proto bird metric 1024 The same configuration for IPv4 works as expected... could you, please, give me some initial orientation regarding the issue?
Hi I think that older kernels (ones that report IPv6 ECMP routes as multiple routes) do not support weight for IPv6. On Jessie, i cannot set IPv6 weighted ECMP even using 'ip route add'. Not sure about newer kernels. BIRD uses older IPv6-ECMP API for route changes (to be compatible with older kernels), so we do not support weight here neither. -- Elen sila lumenn' omentielvo Ondrej 'Santiago' Zajicek (email: santiago@crfreenet.org) OpenPGP encrypted e-mails preferred (KeyID 0x11DEADC3, wwwkeys.pgp.net) "To err is human -- to blame it on a computer is even more so."
Hi Ondrej, thanks a lot for the clarification and your quick response!!! Best Regards Michail Kotsiouros ________________________________ From: Ondrej Zajicek <santiago@crfreenet.org> Sent: Tuesday, January 23, 2018 4:36 PM To: ?????? ?????????? Cc: bird-users@network.cz Subject: Re: IPv6 multipath static routes with different weights and BFD supervision On Mon, Jan 22, 2018 at 03:39:06PM +0000, ?????? ?????????? wrote:
Hi BIRD experts, I face some problems on static routes on IPv6. I try to configure multipath static routes that are supervised with BFD. I use the following conf: route <DEST_IPv6> multipath via <NEXT_HOP1_IPv6> weight 10 bfd via <NEXT_HOP2_IPv6> weight 20 bfd; This does not seem to have the expected behavior since I notice that the static routes are configured with the same priority rather than having the less weighted one with higher priority than the more weighted. When I try to check the routes passed to the kernel I notice this: <DEST_IPv6> via <NEXT_HOP1_IPv6> dev eth3 proto bird metric 1024 <DEST_IPv6> via <NEXT_HOP2_IPv6> dev eth5 proto bird metric 1024 The same configuration for IPv4 works as expected... could you, please, give me some initial orientation regarding the issue?
Hi I think that older kernels (ones that report IPv6 ECMP routes as multiple routes) do not support weight for IPv6. On Jessie, i cannot set IPv6 weighted ECMP even using 'ip route add'. Not sure about newer kernels. BIRD uses older IPv6-ECMP API for route changes (to be compatible with older kernels), so we do not support weight here neither. -- Elen sila lumenn' omentielvo Ondrej 'Santiago' Zajicek (email: santiago@crfreenet.org) OpenPGP encrypted e-mails preferred (KeyID 0x11DEADC3, wwwkeys.pgp.net) "To err is human -- to blame it on a computer is even more so."
participants (3)
-
?????? ?????????? -
Ondrej Zajicek -
Μιχαήλ Κωτσιούρος