-
Notifications
You must be signed in to change notification settings - Fork 520
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
INSTALL: Document kube-proxy settings #391
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Is this the only way? :(
INSTALL.md
Outdated
@@ -230,6 +230,28 @@ Save the file and confirm that the changes have been applied: | |||
kubectl describe configmap -n kube-system aws-auth | |||
``` | |||
|
|||
## kube-proxy settings | |||
By default `kube-proxy` will set the `nf_conntrack_max` kernel parameter to a | |||
small value, overriding [Thar's default |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I have mixed feelings about this wording since it suggests an issue with the default EKS configuration, when there may not really be an issue.
Maybe we need --conntrack-min 0
to mean "apply the new setting but only if it's an increase over the current system value" i.e. an upstream behavior change.
We could also ask EKS to consider changing --conntrack-min
to align with our value.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@bcressey Are you happier with the new wording?
I'll see if I can dig up some more k8s knowledge on this, I wonder if we can set it as part of
Good point, it would be better to say something more along the lines of "...if these values happen to be different". |
In lieu of finding a nicer way to set this, pushed a change to the wording of the instructions. |
I didn't see anything on a quick look over the code that handles this, but I do think we could insert the modification to the DaemonSet during the cluster bootstrapping phase before it starts a node. This would prevent it from making changes between the time we have the Scheduler launch these Pods on a Node and us changing the setting after the fact. |
@bcressey can we merge this? |
The kube-proxy configuration must be updated to prevent it from overriding Thar's sysctl change to nf_conntrack_max. Document the steps to update the configuration in the install steps. Fixes #371 for kernel settings. Signed-off-by: Samuel Mendoza-Jonas <[email protected]>
Rebased |
Issue #, if available:
#371
Description of changes:
The kube-proxy configuration must be updated to prevent it from
overriding Thar's sysctl change to nf_conntrack_max. Document the steps
to update the configuration in the install steps.
Fixes #371 for kernel settings.
Signed-off-by: Samuel Mendoza-Jonas [email protected]
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.