-
Notifications
You must be signed in to change notification settings - Fork 715
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
etcd pod spec should include requests for memory, cpu, etc #2195
Comments
Thanks for noticing. I think we could set it to something sensible by default, that would fit on a default single-master install of 2 (v)CPUs. If we want to get it into the v1.19 release, I could help doing that code-wise. WDYT @fabriziopandini @neolit123? |
what values are you proposing?
even with a |
I confess I had missed that (still can't see it in the code, but I'm sure you're right). And you're also right that critical pods bypass the effect of Edited description. |
actually, i'm mistaken. it is only in master (1.19):
these depend on the cluster size, but i think we can at least set the CPU request for the pod to something like 200m. |
I would like to propose CPU request for the pod to atleast 200m and memory to 100Mi and storage to 1Gi. |
for CPU 200m seems OK, but for memory and storage i think we should leave it to the user to customize. |
Fine that the user can change the setting, but it's sensible to start with some non-zero value. I would go with 100Mi memory and 100Mi disk. |
one option is to gather more comments either from a SIG mailing list discussion and / or during the next SIG CL meeting. @kubernetes/sig-cluster-lifecycle |
ok, it doesn't seem we had any more replies here so let's proceed with @bboreham 's suggestion: mini-guide if a contributor wants to pick this up: modify this function to support update usage of ComponentResources() in the code base to pass make this function: include:
where update unit tests: run:
send the PR. /help |
@neolit123: Please ensure the request meets the requirements listed here. If this request no longer meets these requirements, the label can be removed In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
/assign I would like to help. |
FEATURE REQUEST
The pod spec for etcd does not include any
requests
:https://github.com/kubernetes/kubernetes/blob/db8a88721e587cbc54e7081cb88374e253cde4ed/cmd/kubeadm/app/phases/etcd/local.go#L222-L232
This has the following consequences:
when machine is low on memory, etcd will be a candidate to terminate(Not in 1.19+)kubeadm should add
requests
forcpu
,memory
andephemeral_storage
.It is impossible to pick "correct" values for a component that needs more resources as your cluster gets bigger, but I submit that we can do better by default than zero.
The text was updated successfully, but these errors were encountered: