Issue R3 route traffic to R4 instead of R1. Route 10.10.10.10 toward the WAN is prefered instead of the OSPF LAN DC route. DC Client that try to reach the DC’s 10.10.10.10 address are routed toward the WAN
Why R3 is installing in it RIB and redistributing the wrong routes 10.10.10.10 because it does have a better AD 20.
Workarround If we filter the redistribution that doesn’t help because when the packet arrive to R3 he will still prefer the BGP route.
Below is a very light virtual machine based on Core Linux kernel 4.8 (TinyCore) 26 Mo with network tools like iperf3, tcpdump, net-bridging, iproute2, busybox (httpd), tcpreplay, nmap, openssh.
corelinux1.5.ova
basic commands/directory /etc/sysconfig/tcedir/optional => packages /etc/sysconfig/tcedir/onboot.lst => on boot package to be loaded sudo vi /opt/eth0.sh => change interfaces parameters #configure an interface pkill udhcp => stop dhcp for this interface ifconfig eth0 10.253.106.2 netmask 255.255.255.192 up route add default gw 10.
This post is a memo on how I did the installtion of Kubernetes and Calico on VMs. It’s not some best pactrices in anyway.
I’ve chose VM because I didn’t want to depend on any Cloud infrastructure. I’ve also wanted to understand the network interaction between K8s parts from an infrastructure point of view.
Prerequisite : know how to create VMs on any hypervisors
Steps to deploy K8s : Install 1 ubuntu router with 3 interfaces.
On N7K
https://www.cisco.com/c/en/us/support/docs/switches/nexus-7000-series-switches/116647-technote-product-00.html
=> flanker car avec la commande show hardware internal dev-port-map, il n’y a pas d’asic Clipper, uniquement des flanker
Sample of icmp troubleshooting from the Admin VDC
show module attach module 1 show hardware internal dev-port-map elam asic flanker instance 2 layer2 trigger dbus ipv4 egress if destination-ipv4-address 10.253.108.90 start status elam asic flanker instance 2 layer2 trigger dbus ipv4 ingress if destination-ipv4-address 10.253.108.90 start status On N5K
This is the cheat sheet for the post : https://netmemo.github.io/post/k8s-on-vms-with-calico/
The following post contain raw entry only for reminder purpose.
Bellow are the links I’ve used to understand/did my lab
https://fr.wikipedia.org/wiki/Kubernetes#/media/File:Kubernetes.png https://kubernetes.io/docs/setup/independent/install-kubeadm/ https://kubernetes.io/docs/setup/independent/create-cluster-kubeadm/ https://kubernetes.io/docs/tutorials/k8s101/ https://kubernetes.io/docs/tutorials/k8s201/ https://kubernetes.io/docs/reference/kubectl/cheatsheet/
Join a node/worker to the master
kubeadm join 10.0.1.10:6443 --token d34b9i.v03t2yiozio63cq6 --discovery-token-ca-cert-hash sha256:c21d04ea23790a0bf81cf64118e3a9075ffb63ed90bc697acef5793386e9eb16 Delete a deployment
kubectl delete deployment nginx-deployment-nbo To get the logs of a specific container. -n is to specify the namespace