Description
For the SD-WAN part of my lab, I use 3x C8000V in controller mode as edge routers, which are unmanaged by netlab. I've noticed that recently I can't apply the configuration once the lab is up, because those edges are unreachable.
After some investigations, I've noticed that the issue only occurs after this PR: #3686
I also saw this #3683, which is similar to the problem I'm having. The edge routers no longer have the 10.0.0.15 IP configured.
To Reproduce
groups:
# Unprovisioned devices: configuration will NOT be done by netlab
unprovisioned:
members: [ sdwan-manager, sdwan-controller, sdwan-validator, edge1, edge2, edge3 ]
[...]
nodes:
sdwan-manager:
clab:
type: manager
device: linux
sdwan-controller:
clab:
type: controller
device: linux
sdwan-validator:
clab:
type: validator
device: linux
image: vrnetlab/cisco_sdwan-validator:20.16.1
edge1:
module: [ bgp ]
bgp.as: 65591
bgp.sessions:
ipv4: [ ebgp ]
clab:
type: controller # SD-WAN managed mode, remove the type if using autonomous mode
device: cat8000v
image: vrnetlab/cisco_c8000v:controller-17.15.04c # vrnetlab created two images for c8000v, this is the controller mode
[...]Current workaround
By adding the following to all the c8000v nodes:CLAB_MGMT_PASSTHROUGH: "false"
edge1:
module: [ bgp ]
bgp.as: 65591
bgp.sessions:
ipv4: [ ebgp ]
clab:
type: controller # SD-WAN managed mode, remove the type if using autonomous mode
env:
CLAB_MGMT_PASSTHROUGH: "false"
device: cat8000v
image: vrnetlab/cisco_c8000v:controller-17.15.04c # vrnetlab created two images for c8000v, this is the controller mode
I have now restored my expected behaviour!
Fix
I've also tried recreating the docker image, and it works even better, as I do not need to set the CLAB_MGMT_PASSTHROUGH to false, and the edge devices have the nice management IP address configured and not the 10.0.0.15.
Conclusion
After all this, I realise I don't have much of a point as my problem is solved... I've still created this issue as I will have a look at the documentation to add the workaround if we face an issue with a vrnetlab node not reachable.
This could be related to management Passthrough, in which case you should either rebuild the docker image or set clab.env.CLAB_MGMT_PASSTHROUGH: "false" for the affected node.
Description
For the SD-WAN part of my lab, I use 3x C8000V in controller mode as edge routers, which are unmanaged by netlab. I've noticed that recently I can't apply the configuration once the lab is up, because those edges are unreachable.
After some investigations, I've noticed that the issue only occurs after this PR: #3686
I also saw this #3683, which is similar to the problem I'm having. The edge routers no longer have the 10.0.0.15 IP configured.
To Reproduce
Current workaround
By adding the following to all the c8000v nodes:
CLAB_MGMT_PASSTHROUGH: "false"I have now restored my expected behaviour!
Fix
I've also tried recreating the docker image, and it works even better, as I do not need to set the CLAB_MGMT_PASSTHROUGH to false, and the edge devices have the nice management IP address configured and not the 10.0.0.15.
Conclusion
After all this, I realise I don't have much of a point as my problem is solved... I've still created this issue as I will have a look at the documentation to add the workaround if we face an issue with a vrnetlab node not reachable.
This could be related to management Passthrough, in which case you should either rebuild the docker image or set
clab.env.CLAB_MGMT_PASSTHROUGH: "false"for the affected node.