Options
DMVPN Help
Hello,
I am bashing my head against the wall and really looking for some help with this DMVPN phase three setup. I am not sure what else to do. I have spent hours trying to figure this out with not even as much as a lead. I will explain the setup and problem at a high level. If anyone is interested in helping please let me know and I can send some configs. I am just not sure where else to go.
The Setup (don't fight the scenario
)
OSPF Standard area 55, DMVPN phase three dual hub configuration. Each hub has five spokes. Two of these five spokes are connected to the backbone area 0. Between the spokes in backbone area 0 is a direct connection...in area 0. Clients reside in area 0. Servers reside in area 55. Area 0 has three subnets, one for the direct link, one connected to spoke1, and one connected to spoke2.
The Problem:
If a link on spoke1 and/or spoke2 to one of the hubs goes down, routing converges (very slowly though) and everything is great. When that link comes back up, seemingly random hosts in area zero cannot access seemingly random servers in area 55. All routes, nhrp, and cef look 100% correct. I can even ping the problem hosts/servers from all routers. However, if I reboot a hub or clear an OSPF process on a hub, everything works. And the routes, nhrp, and cef appear to have the exact same entries that were there when it was not working.
After typing this out it 1 - sounds like a bug or 2 - something i believe to be correct is in fact incorrect, such as nhrp, cef, etc looking correct.
Let me know if you think you can help or verify the config.
I am bashing my head against the wall and really looking for some help with this DMVPN phase three setup. I am not sure what else to do. I have spent hours trying to figure this out with not even as much as a lead. I will explain the setup and problem at a high level. If anyone is interested in helping please let me know and I can send some configs. I am just not sure where else to go.
The Setup (don't fight the scenario
)OSPF Standard area 55, DMVPN phase three dual hub configuration. Each hub has five spokes. Two of these five spokes are connected to the backbone area 0. Between the spokes in backbone area 0 is a direct connection...in area 0. Clients reside in area 0. Servers reside in area 55. Area 0 has three subnets, one for the direct link, one connected to spoke1, and one connected to spoke2.
The Problem:
If a link on spoke1 and/or spoke2 to one of the hubs goes down, routing converges (very slowly though) and everything is great. When that link comes back up, seemingly random hosts in area zero cannot access seemingly random servers in area 55. All routes, nhrp, and cef look 100% correct. I can even ping the problem hosts/servers from all routers. However, if I reboot a hub or clear an OSPF process on a hub, everything works. And the routes, nhrp, and cef appear to have the exact same entries that were there when it was not working.
After typing this out it 1 - sounds like a bug or 2 - something i believe to be correct is in fact incorrect, such as nhrp, cef, etc looking correct.
Let me know if you think you can help or verify the config.
Comments
-
Options
gorebrush
Member Posts: 2,743 ■■■■■■■□□□
I haven't looked heavily into dual-hub configurations, but isn't EIGRP the recommended solution over OSPF in DMVPN anyways? Especially in Phase 3.
