Channels
academy
amazon
arm
azure
cabpr
chinese
ci-cd
danish
deutsch
developer
elemental
epinio
espanol
events
extensions
fleet
français
gcp
general
harvester
harvester-dev
hobbyfarm
hypper
japanese
k3d
k3os
k3s
k3s-contributor
kim
kubernetes
kubewarden
lima
logging
longhorn-dev
longhorn-storage
masterclass
mesos
mexico
nederlands
neuvector-security
office-hours
one-point-x
onlinemeetup
onlinetraining
opni
os
ozt
phillydotnet
portugues
rancher-desktop
rancher-extensions
rancher-setup
rancher-wrangler
random
rfed_ara
rio
rke
rke2
russian
s3gw
service-mesh
storage
submariner
supermicro-sixsq
swarm
terraform-controller
terraform-provider-rancher2
terraform-provider-rke
theranchcast
training-0110
training-0124
training-0131
training-0207
training-0214
training-1220
ukranian
v16-v21-migration
vsphere
windows
Title
b
big-judge-33880
02/13/2023, 11:30 AMIs there any guidance on pitfalls related to reconfiguring the networks in an existing harvester cluster? Currently all machines on vlan networks are connected to the management cluster network (due to initial hardware constraints), but with more network interfaces I’d like to fan out and use separate VM and storage networks
g
great-bear-19718
02/13/2023, 10:08 PMcould you please elaborate? are you planning to add more VM networks and migrate workload VM's to use the new VM networks?
b
big-judge-33880
02/14/2023, 8:24 AMI’m currently running the management network with no vlan set, and would like to apply a vlan id to it. Ideally I’d want to move this to a new network, too, but I assume that might be more difficult in practice.
Then I’d like to move all the VMs currently using the management network over to a new cluster network for VMs, while adding a separate storage network for longhorn to use.
(aka break things, probably)
I’d also want to change the physical interfaces used for the management bond, so I’m thinking of something like this admittedly naive plan:
1. Making changes in /oem configuration for management bond to use 1Gbit interfaces (currently running on a bond of 10Gbit interfaces) - same IP addresses as before, but now vlan-tagged
2. Shut down and restart all nodes
3. Create two cluster network for vms and storage (VM network to reuse interfaces previously used by management)
4. Shut down all VMs
5. Attach all VMs to new networks connected to new vm cluster network
6. Set up storage network according to docs
7. Start all VMs and cross fingers
g
great-bear-19718
02/14/2023, 10:31 PMchanges to /oem/harvester.config will not be applied to the underlying interfaces. that is more of a place holder and is used to generated /oem/99_ custom.yaml which has the OS configuration instructions
b
big-judge-33880
02/14/2023, 10:50 PMYeah, thank you. I assumed that since harvester.conf contains the initial setup of the cluster from an older version.
Am I correct in assuming I can simply replace the interface names associated with the management network in 99_custom.yaml if I ensure the the same vlan and aggregation settings are available on the switch connected to the new interfaces? As far as I can tell /etc/sysconfig/network/ gets overwritten on boot, so no need to clean out the old interface configs?
g
great-bear-19718
02/14/2023, 10:51 PMyeah 99_custom.yaml is what is applied, in there they will be a setup script for the interfaces which applies the vlan id when creating the br and bo interfaces
✅ 1
b
big-judge-33880
02/14/2023, 11:02 PMThanks a lot for your input!