How to boot a Ubuntu vm into single user mode via harvester UI vm console shortcut key? Assuming such VM could not be logged into via all means, and need to be able to boot into single user mode to update

ubuntu

user password, and allow console login..

Are there other supported options for VM backups (like Veeam) with Harvester or we only have NFS and S3 option? https://docs.harvesterhci.io/v1.1/vm/backup-restore Above doc shows "Available as of v0.3.0"... so just wondering if there are more options now in 1.1?

hello, I am trying to rename a network interface in harvester, does anyone know of a best practice/documented method for doing so?

Hello! Is it possible to turn off the monitoring to save on resources? Before I do that though if I can, should the cluster be writing 2.5MB/s to disk constantly when idle with no vms?

can someone provide the steps to add Harvester CSI to a custom rancher cluster? looks like the helm chart is looking for cloud-config folder, and there should be a file cloud-provider-config in that folder. how can i generate that file? i know it's all done automatically if you create a cluster in rancher w/ harvester as a cloud provider, but i'm doing a custom cluster so i can add windows nodes.

Hello, During installation of the first Harvester node, came across this issue (screenshot) Any idea why this happens ?

Hello, after (somehow) I managed to make harvester start, I see this issue

image.png

Can someone advise on this ?

shouldn’t this be public image ? Pull if also failing for me locally

image.png

Anyone seen this issue ?

Normal   Pulling                 39m (x2 over 40m)     kubelet            Pulling image "rancher/harvester-cluster-repo:v1.1.1"

Normal   BackOff                 39m (x3 over 40m)     kubelet            Back-off pulling image "rancher/harvester-cluster-repo:v1.1.1"

Warning  Failed                  39m (x3 over 40m)     kubelet            Error: ImagePullBackOff

Normal   Pulling                 37m (x4 over 39m)     kubelet            Pulling image "rancher/harvester-cluster-repo:v1.1.1"

Warning  Failed                  37m (x4 over 39m)     kubelet            Failed to pull image "rancher/harvester-cluster-repo:v1.1.1": rpc error: code = Unknown desc = failed to pull and unpack image "<http://docker.io/rancher/harvester-cluster-repo:v1.1.1|docker.io/rancher/harvester-cluster-repo:v1.1.1>": failed to resolve reference "<http://docker.io/rancher/harvester-cluster-repo:v1.1.1|docker.io/rancher/harvester-cluster-repo:v1.1.1>": pull access denied, repository does not exist or may require authorization: server message: insufficient_scope: authorization failed

Warning  Failed                  37m (x4 over 39m)     kubelet            Error: ErrImagePull

Warning  Failed                  37m (x6 over 38m)     kubelet            Error: ImagePullBackOff

Normal   BackOff                 4m6s (x154 over 38m)  kubelet            Back-off pulling image "rancher/harvester-cluster-repo:v1.1.1"

Hello, I registered Harvester into Rancher but not able to configure a network, in order to select it when creating node template. Can someone advise on that ?

I created vm network and was able to select it in the node template. However I’m facing this issue: I’m creating a new cluster, (1 master 1 worker) and selecting the corresponding node templates Then I see the cluster provisioning, 2 new vms are created, and they are in running state. However they never complete and register to the cluster… the cluster hangs in “Provisioning” I then noticed that the Storage of the vms is in “degraded” state. Any Advise on that ?

image.png

I'm making good progress toward adopting Harvester. Now I'm assessing the basic capabilities by creating a single-node K3s cluster through Rancher Server UI. I am stuck trying to diagnose a pending state. Can you direct me to find the cause? Cluster status:

waiting for viable init node

Conditions log:

[Disconnected] Cluster agent is not connected

How does IPAM inside harvester work? If not depending on DHCP for VM IP assignment?

Is it possible to clean up the host's failed storage allocation cards in the UI? They are artifacts from attempting to force format a disk that was considered in-use because it had a present LVM volume. Now I have added the same disk successfully, but the error cards still remain in the UI.

Hi, I think I have a pretty basic issue here installing harvester, https://github.com/harvester/harvester/issues/3672, can anyone help?

we have a kubernetes cluster running on harvester, created via an external rancher. we had a brief outage earlier today, and harvester seems to have rebooted all relevant vms. however, since the vms are all assigned ips via dhcp, and we had a fairly short lease time configured, they all changed ip addresses. this was reported correctly to harvester, which displays the new ip addresses, but the kubernetes cluster in rancher still seems to expect the old addresses for the nodes. is there anything we can do to flush those?

I have gone through the Harvester docs, but it's unclear how I can setup networking to achieve the following: I have a server with single interface. LB VM which need to be accessed from the internet and several VMs behind the LB whose IP need to be on a private subnet (these VM need outgoing connection, but not incoming). Anyone know how this can be setup?

I expected to see the Longhorn storage classes I created in Harvester available in the K3s cluster I created in Rancher, but only the default class

<http://rancher.io/local-path|rancher.io/local-path>

appears. There must be another step. I see this in StorageClass doc:

Different Longhorn StorageClasses might map to replica policies, or to node schedule policies, or disk schedule policies determined by the cluster administrators

i'm trying to get a custom rke2 or k3s cluster working on-top of harvester. but i'm having issues getting the cloud-provider working. i followed: https://docs.harvesterhci.io/v1.1/rancher/cloud-provider if i deploy a 1.24.4 RKE2 cluster with Harvester for the cloud provider, the cluster never finishes spinning up. if i try to do a custom RKE2/K3S cluster with no provider, then add the Harvester-Cloud-provider later, i get this error: Error getting instance metadata for node addresses: Get

"<https://192.168.5.100/k8s/clusters/local/apis/kubevirt.io/v1/namespaces/k8s/virtualmachines/k3s-m1>": x509: certificate signed by unknown authority

i imagine because my harvester cluster has a self-signed cert. any ideas on how to fix this? seems like my best bet is a custom-cluster for the provider, but i have no idea how to get the k3s or rke2 guest k8s cluster to trust harvester vip

Is there away to setup a NAT on harvester so that I can have several VMs on a private subnet (VMs would have static IPs)?

How does Harvester/Kubevirt load balancing VMs (DRS) around multiple nodes in the same cluster? With 3-node (very powerful physical hosts) empty Harvester cluster, I deployed 3 small VMs, and it seems it deployed these 3 VMs on the same physical host. Is this just randomness? Or is it because it would prefer packing up one node first, before deploying new VMs on other nodes in the cluster?

Is it necessary to enforce a NetworkPolicy, e.g., by Namespace, in order to isolate a Harvester VM network's layer 2? I'm trying to understand and explain why a DHCP server on a VM network of type

L2VlanNetwork

is answering lease requests originating from the attached LAN of the Harvester host's metal NIC. The VM network uses the default mgmt cluster network.

Why in Harvester, I got this for new VM running on top of it?

# dmidecode -s system-serial-number
Not Specified
# dmidecode -s system-product-nme
None

I get this error when I try to build an RKE2 cluster with Harvester as cloud provider.

waiting for viable init node

The nodes are launched, but the console only shows a cloud-init error until the pool is eventually deleted. Wash, rinse, repeat. How can I diagnose the node condition that Rancher is waiting for?

When creating an RKE2 cluster with the Harvester Cloud Provider, how can I influence the Harvester storage class that is selected for guest cluster workload PVCs? In the guest cluster, the visible storage class is "harvester", and the driver always picks the Harvester storage class "longhorn" which was the default when the Harvester was installed, but is not the current default storage class.

Hi, gents