Hi, I just received my RK1 and so i have replaced my node 1 cm-4 , I followed the guide to set up the RK1 with ubuntu and kubernetes and it seems to work except for one thing. All of the other nodes i can access ssh using the node#.local but not the new one. my nodes are called turing-node-1 thru 4. I'm wondering what in the network config might cause that ?

How did you install K8s on the RK1?, eg did you ssh into ubuntu.local?, can you still do this?, if so have you changed the host name from the default “ubuntu” to the new host name eg “turing-node-1” , you can change the hostname by either using “hostnamectl set-hostname ” or by editing the /etc/hostname file directly.. eg sudo nano /etc/hostname

I disd the sudo nano /etc/hostname

what about ssh user@ubuntu.local, if that works I suggest a reboot of the node

when logged in it shows the correct host name.

are you using a static IP, or DHCP either dynamic or reserved

static and its reserved on my main router

ok, you have created a host entry in your DHCP server with a reserved IP address?, you will need to delete or update the MAC address associated to the RK1

I notice my /etc/resolv.conf on the other nodes is different

if you use the arp utility from fedora do you see the host in the list

what is the foramt of the command ?

im on ubuntu but think arp -a should work

123 = node 3, 122=node 2 and 121= node1

mmm mDNS issue, Im not familiar with this, but can you clear/flush the DNS cache

let me see

also I’d delete the message above with your MAC address… You can’t be to careful :-).

I don't have that option on the router

oh ok, you could manually add the hosts to your hosts file on your fedora and nodes, a little bit of a pain, but would work around not having access to DNS configuration

Yeah I have put all of the nodes into the hosts file of each node I can try that, it's just that it works auotmatically for the other nodes I think its a mulicast issue

ok, good luck

I noticed in the /etc/systemd/resolved.conf that the commented out mdns=no on node 1 and yes on the others, trying that

Just done a little reading on multicast DNS, very cool, so yes hopefully the isssue

found it on the other nodes they are running avahi deamon