https://rancher.com/ logo
Join Slack
Powered by
This message was deleted.
# harvester
a
This message was deleted.
w
it's the cluster VIP address which you configured when setting up your first node
r
so i redid the install and used the mgmt ip from server 1 as the mgmt url for server 2 but im still getting "unavailable" very strange
w
presumably server 2 can ping server 1? and you can log in to the mgmt url from another machine?
r
yes to both
im upgrading both from 1.3.1 to .2 now to see if that fixes it
w
if server 1 works i'd upgrade that first then install server 2 using 1.3.2 to join server 1 - you can't join two installed single servers post-install
✅ 1
r
QQ: does the mgmt url need to have protocol? 
x.x.x.x
vs 
<https://x.x.x.x>
? not sure if that matters
nvm - when i type in the ip and press enter the installer is actually pinging now. must be a new 1.3.2 feature i would have deleted the Q but apparently deletes are disabled in this slack /shrug
w
it's been a while since i installed Harvester - I started with 1.0.0 then upgraded through versions to 1.3.2 - but I know the installer has changed & improved
r
woof. exact same thing. this is so dumb... literally just says mgmt url unavailable from node 2 i can 
curl -k <https://harvester.local/ping>
and it resolves fine (this is node 1 mgmt url)
w
is the cluster token correct?
r
yeah its correct
w
it's going to be something network-related (dns, routing), time sync, etc.
is this failing during install of server 2 or after it's finished installing?
r
install finishes but it never finds the mgmt url... Status: NotReady... ik this can take a while but it just sits like this forever only error i can tell is from /var/log/console.log ~~ 
couldnt get current server API group list: GET <https://127.0.0.1:6443> was refused
hard to know if that is even related since i see a similar error on the master node but with port 8080
w
it can take some minutes for things to sync but not hours/days
r
Copy code
harvester01:~ # kubectl get nodes -n harvester-system
NAME          STATUS     ROLES                       AGE    VERSION
harvester01   Ready      control-plane,etcd,master   108d   v1.28.12+rke2r1
harvester02   NotReady   <none>                      69m    v1.28.12+rke2r1
This is interesting... so node 1 is actually seeing node 2 but not finishing setting it up
@witty-jelly-95845 i finally got this working btw. so dumb. where the installation asks for mgmt information... it reads like an IP address. if you put in an IP it fails, dns name it fails because of certs, but if you turn the IP into a URL it works... so dumb. im gonna send them an issue/ PR to update the docs
👏 1
w
Yes that is dumb!
Maybe something has changed since earlier versions but I'm sure I used IP address to join my second node to first as I didn't have DNS then.
3 Views
Open in Slack
PreviousNext
Powered by