cluster-lab
cluster-lab copied to clipboard
Is it possible for Cluster Lab on wlan0 instead of eth0?
Hello, I've followed the instructions in this Hypriot Blog Post. Could that be setup Cluster Lab using Pi 3's wifi chip instead of using eth0 (Ethernet)?
I tried to edit /etc/cluster-lab/cluster.conf manually. When I cluster-lab stop and then start. It shows me:
Got more than one IP address on wlan0.200: 192.168.200.1
Is that a bad idea to build a cluster using wlan? If not, how to config Cluster Lab on the Pi 3? Thanks!
@firecyberice Can you guide @oktak through this?
@oktak You just need to change these three parameters:
- https://github.com/hypriot/cluster-lab/blob/master/package/etc/cluster-lab/cluster.conf#L9
- https://github.com/hypriot/cluster-lab/blob/master/package/etc/cluster-lab/cluster.conf#L13 (wlan does not support vlan)
- https://github.com/hypriot/cluster-lab/blob/master/package/etc/cluster-lab/cluster.conf#L20 (assuming that your wifi network already has a dhcp server)
Thanks for @firecyberice help. I got cluster-lab health all [PASS] now :D
Here is my journey:
Before I open an issue here, I used --clusterlab true flag (with ssid, password, hostname together) to flash the microSD Card for the Pi 3.
I found its default is using eth0. I have not much LAN cable in hands, so I tried to modify /etc/cluster-lab/cluster.conf manually to "wlan0". Doing this I got error message above.
I tried to follow the steps and changes those parameters. However I got 'No route to Host' when I connect the Pi 3 via SSH. I tried several times for different values.
Finally, I start it over, I flash the microsSD again, without --clusterlab true flag.
- Boot HypriotOS Cluster Lab on Pi 3 and SSH connect to it.
- I change value of INTERFACE in
/etc/cluster-lab/cluster.confto "wlan0" (I changed only this one for my "incremental test" purpose), and save it. - I use
VERBOSE=true cluster-lab startto see what happens. - Oh yeah! almost all [PASS]. I can access at http://{IP of the node}:8500 to see the Consul web interface.
- To see if there are any [FAIL]s, I typed
cluster-lab health, this time, really all [PASS]. Thanks so much :)
I suspected that --clusterlab true flag would go for some config assuming eth0 in use for the first boot. Also, it appeared that it also appends some docker related values in device-init.yaml (this point I am not sure, could be I wrongly added to it...oops)