https://rancher.com/ logo
Join Slack
Powered by
This message was deleted.
# rke2
a
This message was deleted.
a
I run it on cilium, rke2 docs are clear, except for one thing, that you need to re-create cluster
c
You shouldn’t change CNI after the cluster is up, if that’s what you mean.
a
There are probably ways to migrate, but hard to do
c
the values file for cilium’s helm chart is at https://github.com/cilium/cilium/blob/v1.13.4/install/kubernetes/cilium/values.yaml - you should be able to poke through that what options are available.
👍 1
m
Ow, so I would need to recreate the cluster. Got it thanks!
c
You would need to recreate the cluster to change from another CNI to cilium. Not to change the cilium configuration.
m
I was about to reset the cluster. I am already running Cilium, we only need to enable other features. How do we do that without resetting the cluster?
a
check values
m
Sorry, I’m still new here. What 
check values
?
a
in the file @creamy-pencil-82913 provided
values.yaml
I'm looking for my example
hubble:
enabled: true
metrics:
enabled:
- dns:query;ignoreAAAA
- drop
- tcp
- flow
- icmp
- http
relay:
enabled: true
ui:
enabled: true
👍 1
m
Awesome, looking at the same thing. Just one follow-up question though, how do we know which version of the code was deployed on the current cluster? The shared link is 
v1.13.4
a
cilium daemonset
has image tag
something like rancher/mirrored-cilium-cilium:v1.13.2
👍 1
m
nice, I found it
Copy code
apply-sysctl-overwrites:
    Image:      rancher/mirrored-cilium-cilium:v1.13.2
    Port:       <none>
    Host Port:  <none>

...

   install-cni-binaries:
    Image:      rancher/mirrored-cilium-cilium:v1.13.2
    Port:       <none>
    Host Port:  <none>
    Command:
      /install-plugin.sh
Ok, let’s say we added the changes. How do we apply the change to the cluster?
a
ah yes, I don't think they apply themselves right away, you need to restart the daemonset sometimes, what I normally do, I don't know if it's the right way or proper way to do but I delete the cilium pod, kind of hard resetting the system
m
Ow, I’ll go with that for now. Thanks!
c
deploying an updated HelmChartConfig with new values should cause the chart to update and update the pod config
m
You mean, that’s automatically detected by RKE2 or do we need to run a command? EDIT: Ow, Just checked helm: 
rke2-cilium                     	kube-system	2       	2023-08-01 05:05:08.052874208 +0000 UTC	deployed	rke2-cilium-1.13.200                        	1.13.2
That means we can run 
helm update rke2-cilium
?
c
no. update the valuesContent in the HelmChartConfig you deployed
just edit 
/var/lib/rancher/rke2/server/manifests/rke2-cilium-config.yaml
, it’ll update automatically
you should see the chart get redeployed with the updated values
m
Awesome! Thanks for the confirmation.
c
assuming everything is valid
m
Just a follow-up question. How long do we need to wait before RKE2 actually updates with the edited file?
a
Immediately, but if you update configuration it’s not guarantee for service restart
m
hmm, a few hours passed now and did not update. even restarted the rollout of cilium-operator and ds.
a
Now you see the reason I wrote you earlier?
m
yeah, looks like we need to hard reset things for the changes to take effect
w
Hard reset means? Just curious to know what procedure you have followed?
936 Views
Open in Slack
PreviousNext
Powered by