Skip to content
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

Documentation: 3 or more Masters into Regions with only 2 AZ #3389

Closed
nilocbb opened this issue Sep 15, 2017 · 4 comments
Closed

Documentation: 3 or more Masters into Regions with only 2 AZ #3389

nilocbb opened this issue Sep 15, 2017 · 4 comments
Labels
area/documentation good first issue Denotes an issue ready for a new contributor, according to the "help wanted" guidelines.

Comments

@nilocbb
Copy link

nilocbb commented Sep 15, 2017

Heres my problem:
I only have a Region with 2 AZ. When I created a cluster with 3 masters split into 2 AZ I ran into connection lost every now and then.

kubectl get nodes sometimes gets the nodes sometimes not.

I used it in a Private subnet.

I resolve this using 3 masters in only one AZ and nodes split between the other two.

Maybe we should Doc this to alert when creating a cluster (best practice).

@jatan28
Copy link
Contributor

jatan28 commented Sep 30, 2017

I can work on this, also would it make sense to support having 3 masters balanced across 2 AZs in the regions that has only 2AZs.

@slongstreet
Copy link

I would categorize this is a "known issue", not a best practice. Putting all 3 masters in a single AZ means it isn't highly available -- this is just a workaround until kops supports specifying only two AZs for the master zones.

@chrislovecnm
Copy link
Contributor

Related: #3503

@chrislovecnm
Copy link
Contributor

We just want to make sure that we have (N-1)/2 spread across two AZ to have HA. Creating an instancegroup type for just etcd, may be best, but I am open to other options.

k8s-github-robot pushed a commit that referenced this issue Oct 5, 2017
Automatic merge from submit-queue.

Added documentation on best practices for creating clusters with HA m…

Fix #3389
@rifelpet rifelpet added good first issue Denotes an issue ready for a new contributor, according to the "help wanted" guidelines. and removed good-starter-issue labels Apr 18, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
area/documentation good first issue Denotes an issue ready for a new contributor, according to the "help wanted" guidelines.
Projects
None yet
Development

No branches or pull requests

5 participants