Skip to main content
Announcements
Qlik Connect 2024! Seize endless possibilities! LEARN MORE
cancel
Showing results for 
Search instead for 
Did you mean: 
Sandra_Ospina
Partner - Contributor II
Partner - Contributor II

QSE on EKS pods keep on showing as “pending”

Hi all,

I’m Installing QS Enterprise on EKS following the Qlik help documentation. However, many of my pods keep on showing as “pending”.  When I try to connect via a web browser, I get  an empty page.

clipboard_image_4.png

I checked the logs and I got the following message:

“0/3 nodes are available: 3 Insufficient pods” 

Pods:

clipboard_image_0.png

Nodes:

clipboard_image_1.png

Any suggestions? Does anyone have guidance for setting this up using EKS?

Thanks!

Labels (3)
1 Solution

Accepted Solutions
ThiebaudS
Partner - Creator II
Partner - Creator II

Hi again,

In EKS, if you are using the default Amazon VPC CNI plugin for Kubernetes, whenever you deploy a Pod on a worker node, EKS creates a new IP address from VPC subnet and attach it to the instance.

Therefore, there is a max pods limit for each EC2 instance type, depending on the maximum supported Network Interfaces for this instance type, and the IPv4 Addresses per Interface.

You can find these information for each instance type here: https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/using-eni.html#AvailableIpPerENI

Then the formula to find the max pods limit is:

Max Pods = ( Maximum supported Network Interfaces for instance type ) * ( IPv4 Addresses per Interface ) - 1

In your case, with t2.medium instances, you have a limit of  3*6 -1 = 17 pods per instance (including K8S system pods).

QSEoK is using more than 60 pods so the answer is yes, the instance type is too small.

I suggest you to try with at least 2 t3.xlarge.

Best regards,

Thiebaud

 

 

View solution in original post

4 Replies
ThiebaudS
Partner - Creator II
Partner - Creator II

Hi Sandrar,

What are the EC2 instance type of your worker nodes ?

Best regards,

Thiebaud

Sandra_Ospina
Partner - Contributor II
Partner - Contributor II
Author

Hi Thiebaud,

The Instance type is t2.medium

Is it too low? I can try to increase the instances to t2.xlarge or higher. Please advise.

Thank you!

ThiebaudS
Partner - Creator II
Partner - Creator II

Hi again,

In EKS, if you are using the default Amazon VPC CNI plugin for Kubernetes, whenever you deploy a Pod on a worker node, EKS creates a new IP address from VPC subnet and attach it to the instance.

Therefore, there is a max pods limit for each EC2 instance type, depending on the maximum supported Network Interfaces for this instance type, and the IPv4 Addresses per Interface.

You can find these information for each instance type here: https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/using-eni.html#AvailableIpPerENI

Then the formula to find the max pods limit is:

Max Pods = ( Maximum supported Network Interfaces for instance type ) * ( IPv4 Addresses per Interface ) - 1

In your case, with t2.medium instances, you have a limit of  3*6 -1 = 17 pods per instance (including K8S system pods).

QSEoK is using more than 60 pods so the answer is yes, the instance type is too small.

I suggest you to try with at least 2 t3.xlarge.

Best regards,

Thiebaud

 

 

Sandra_Ospina
Partner - Contributor II
Partner - Contributor II
Author

Thiebaud,

I increased to t3.xlarge, and now all of my pods are running.

Thank you!