https://rancher.com/ logo
Join Slack
Powered by
This message was deleted.
# harvester
a
This message was deleted.
t
Looks like you have the networks flipped. In the middle image ( creating the vm ) you have have the management network first. this will connect to eth0 on the vm. Also does your main-network have an physical Nic attached on the host?
p
Yes, the hosts have a physical Nic for that cluster network. That network uses 10.5.106.X So regarding the middle image, I think that's how it should be - get the mgmt IP through DHCP and then I have a static IP on eth1 for me to access from outside the cluster
t
ok, that is intended?
p
The hosts can reach the VMs (though the VMs cannot reach the hosts, I think that's another issue)
t
did you change anything on the mgnt network?
p
I previously didn't have to have the mgmt IP. But currently my LBs in Harvester gave up on working, so I wanted to try that. And then this
Didn't change anything I can remember, no
t
let’s start from the bottom, how many nics are on the host(s)?
p
2. One for mgmt and the other for the main network vlan. Though even with the latter off, the host can ping the vlan IP of the VM
Though, to be honest, I don't really care if the VM gets a mgmt IP or is even part of the mgmt network. I just want my Load Balancers to work again 😅
t
So remember NIC --> “network” --> VM. so does your host need 2 nics? You can also vlan tag the host nics. my setup is everything over a single nic, mgnt, and a single “network” using the mgnt nic.
p
You're right, I don't need the 2nd NIC. If I traceroute host to VM, it goes from host to gateway, to VM on other vlan
t
did you create a “virtual machine network” with an untagged vlan? Here is what mine looks like.
p
No, I believe mine is tagged. It was working fine till like 2 days ago
t
are you actually using vlans?
p
I am thankful that the cluster hasn't progressed to the point where a full reinstall isn't out of the question
I'm not a networking person, but the network guy did send it as a vlan
t
I would check with him to see if anything changed. Also can you ssh into the host? I would start with the host networking and see if it is working correctly. Then the “vm network”, and the the VM’s itself. so 3 layers.
p
I have the support bundle, I might try to find if there's something abnormal or which could have changed
supportbundle_afbe44ef-9329-41ce-85ea-3951ec79ce18_2025-02-05T10-13-32Z.zip
t
what version of harvester are you using?
p
I will check with him. So far, I can go host -> VM and my laptop -> VM/Host. It's just VM->Host which doesn't work, the traceroute doesn't even reach the gateway of that vlan. Load Balancers in Harvester can no longer be pinged, and time out if accessed.
It's 1.4.0
Again, it had been working fine for some 60 days and only just now decided to break
t
1.4.1 is out with fixes.. maybe worth a try. did you reboot. lol?
p
You know what
Damn good point, lemme try that
t
lmao!
p
I don't deserve to work in IT if I didn't think of reboot first, debug second
haha
t
lmao…
p
Oh. My. God.
t
drum roll……
p
If ever I have the pleasure of meeting you, I'm buying you a drink.
sigh
Thanks a ton haha. Reboot and it works fiiiiiiine now
t
Sweet. I am in santa clara this week, denver next, and annapolis the rest of the time.
lol NIOCE!
Open in Slack
PreviousNext
Powered by