By the way, just got a CloudCoreRouter 1036-8G-2S+ to handle my home’s multiple 2x 1 Gbps lines, everything’s mostly working fine. With the exception of some minor issues with RouterOS and then some slightly more troublesome issues. I’m running RouterOS 6.10 and on the newest firmware.
I have a GRE tunnel up to a remote endpoint, and add a static route to reach the endpoint over the GRE tunnel. I can ping endpoint, now the problem comes when I establish a BGP session with the endpoint, I get the routes fine, but then in IP => Routes => Nexthops, the endpoint’s nexthop is my WAN gateway which is wrong. Anyone has any idea on how to fix it? The address I have assigned my GRE tunnel is on a different segment than on the remote endpoint’s.
I find that my CCR1036-8G-2S+ has a nasty habit of locking up the user interface, by user interface, I mean telnet, ssh, winbox, http. When I log in to the winbox and http interface, in all the menus, all the menus are empty and without any items. And then when I try to log in to telnet / ssh and then print anything from CLI menus, it takes many many minutes to print any output. During the time when the interface locks up, routing’s still working fine, just that I can’t change any settings unless I reboot the router. (Kinda a pain to do when you have multiple users sharing it)
Just reporting what I have experienced, and just curious if anyone has encountered any of these issues and if any ideas how to fix the above following issues?
Here’s the snippets of configuration for the BGP peer. The other endpoint is not controlled by me, and is not a Routerboard / RouterOS, but if I remember correctly, either a Quagga / BIRD server.
/ip address
add address=172.23.253.1/32 interface=gre-tunnel3-siska network=172.23.253.1
/ip route
add distance=1 dst-address=172.23.4.131/32 gateway=gre-tunnel3-siska
/routing bgp instance
add as=76136 client-to-client-reflection=no name=NET-LEROY-DN42 out-filter=
DN42-in
/routing bgp peer
add address-families=ip,l2vpn,vpnv4 comment=“siska DN42” in-filter=DN42-in
instance=NET-LEROY-DN42 multihop=yes name=siska-dn42-us out-filter=
DN42-in remote-address=172.23.4.131 remote-as=64542 route-reflect=yes
ttl=defaultWhen looking at the announced routes in the route table from BGP. The gateway is correct: 172.23.4.131, but then the nexthop is recursive via (WAN IP) ether1-gateway.
Hi normis, just an update.
I have just added another peer using the following commands below. And the routes announced by the peer have the correct gateway of 172.22.41.232, but nexthop recursive via my WAN gateway - ether1-gateway.
Have you tried nexthop-choice=propagate which will: “try to propagate further the nexthop received; i.e. if the route has BGP NEXT_HOP attribute, then use it as the nexthop, otherwise fall back to the default case”?
“set-in-nexthop (IP address;) set gateway value to the specific IP address[es]. Valid only in incoming filters”
There is also an ability to set the nexthop in a route filter.
chain=DN42-in … set-in-nexthop=1.2.3.4
Hmm. joshaven, tried specifying nexthop explicity using a passthrough in filter. But it too results in a (specified nexthop) - recursive via WAN gateway (ether1-gateway)
Using nexthop-choice=propogate seems to have no effect too. So far three out of my five peers seem to not have this problem.
I’m a bit lost here, wondering if its my configuration or something I did wrong.
Hi joshaven, thanks for your ideas previously.
Just to update this thread with the response that I got form a MikroTik staff from the support email I sent.
On a second note, seems like there is still issues with the user interface locking up necessitating a reboot.. pokes normis for a fix.
