You need to configure a Google Kubernetes Engine (GKE) cluster. The initial deployment should have 5 nodes with the potential to scale to 10 nodes. The maximum number of Pods per node is 8. The number of services could grow from 100 to up to 1024. How should you design the IP schema to optimally meet this requirement?
To configure a Google Kubernetes Engine (GKE) cluster with the given requirements, the IP schema design should be as follows: First, a /28 primary IP address range for the node IP addresses is appropriate as it can handle up to 16 addresses (14 usable IP addresses for nodes, considering the reserves). Second, the Pods' secondary IP range needs to be /24, since a /24 can handle 256 IP addresses, which comfortably accommodates the potential maximum of 80 Pods (8 Pods x 10 nodes), and /24 is also the smallest subnet range for Pods in GKE according to the documentation. Lastly, a /22 secondary IP range for the Services should be used as it provides 1024 IP addresses, which matches the upper limit of the services (1024). Therefore, the IP schema requires /28 for nodes, /24 for Pods, and /22 for Services.
/22 has only 1022 usable IP addresses, looks like B is the correct answer.
yes, we need pay attention, was required 1024 valid address, cann't be A. Correct is B
secondary subnets can use all available space /22 = 1024. Answer is D.
Secondary subnets are subject to the same limitations of primary subnets. A /22 would not suffice in this case even for secondary subnets, must be /21.
They are not subject to that limitation. Source: https://cloud.google.com/kubernetes-engine/docs/concepts/alias-ips#cluster_sizing_secondary_range_svcs Note how the table for services on secondary IP range says /22 allows for the full 1024 services. Meanwhile if you scroll up to the primary range for nodes, it says a /22 would only allow for 1020 or 1019 nodes, because primary ranges do have that constraint. D is the only possible answer, also because /24 is the smallest possible pod range (/25 and /28 not possible), mentioned in the same documentation page.
Agree with your answer
Vote for D 10 nodes = /28 can max 12 nodes 8 Pods per node = 80 Pods = /24 max 110 Pods 1024 services = /22 max 1024 services source: https://cloud.google.com/kubernetes-engine/docs/concepts/alias-ips
Not correct. IP Address for Services: /22 => 2^10 = 1024, but NOT ALL 1024 IPs are avaiables, there are some IPs reserved for GCP: 0.0: Network address. 0.1: Reserved by AWS for the VPC router. 0.2: Reserved by AWS. The IP address of the DNS server is the base of the VPC network range plus two. ... 0.3: Reserved by AWS for future use. 0.255: Network broadcast address. B correct answer (/21 address space for Services)
I'm sorry. Correct Anwser D, as exposed by Thornadoo.
This is a tricky question. Keep in mind that Nodes use Primary IP ranges whereas Pods and Services use Secondary ranges. For secondary, as you can use all of the IP addresses unlike the Primary where 4 x IPs are reserved. E.g. Max hosts in /24 for Node is 252 vs. 256 for pods and services. If we gather the above logic: 1) 10 nodes = /28 2) 80 pods = /24 (Reason is /24 is the smallest possible Pod) 3) 1024 services = /22 (Can use all of 1024 IPs) https://cloud.google.com/kubernetes-engine/docs/concepts/alias-ips#cluster_sizing_secondary_range_pods
It says /24 is the is the smallest possible Pod for user-managed... we're using GKE, not user-managed. ;)
Answer is A
In addition Answer D say 'Pads' not 'Pods' so it's wrong already. I wonder if the exam question show it like that. Dead giveaway for a wrong answer!
It should be a typo, they don't want to examine our English skills
The question does not specify, but it's irrelevant anyway, since the smallest range for GKE-managed is even larger at /21. The answers all mention pod ranges of /24 or smaller already. If anything, that's even more reason for D to be the correct answer - both /25 and /28 are not possible ranges for pods.
Agree with you!
8 pod per node = 16 ips per node * 10 nodes = 160 ips > /24
Plus seconday ranges don't have reserved ips so we are fine with a 1024 = /22
All the correct Maths responds to B
Actually it must be D, as the minimum POD IP-Range is /24. This eliminates all other possibilities, right? Source: https://cloud.google.com/kubernetes-engine/docs/concepts/alias-ips?hl=de#cluster_sizing_secondary_range_pods
vote for B, below the explanation: max node is 10 = /28 max pod is 80 = /25 max service 1024 = /21, can't use /22 because only 1022 usable ip. so the answer is B
vote for B, below the explanation: max node is 10 = /28 max pod is 80 = /25 max service 1024 = /21, can't use /22 because only 1022 usable ip.
10 nodes = /28 (14 maximum usable IP address) 8 Pods per node = 80 Pods. But Number of IP addresses=16*10=160 = /24 (254 maximum usable IP address). Please refer: https://cloud.google.com/kubernetes-engine/docs/how-to/flexible-pod-cidr 1024 services = /21 (2046 maximum usable IP address) (/22 is not applicable as 1022 maximum usable IP address). https://www.subnet-calculator.com/cidr.php But none of the above options are fully correct. So, I will go for D
Option B specifies a /21 secondary IP range for Services, which is larger than necessary for the number of services stated in the question. Option C specifies a /28 secondary IP range for Pods, which is smaller than the minimum IP range required for Pods specified by GKE. Option D specifies a /24 secondary IP range for Pods, which violates the maximum number of Pods per node stated in the question.
B is correct. Remember 1024 is the total IP in the mask /22. But the free IP to use is 1022. because of 2 more IPs, the correct is /21.
The correct answer should be D: When we have 10 nodes, we need 14 IP addresses, the first 2 and the last 2 IP addresses will be reserved, so /28 which gives 12 IP addresses. https://cloud.google.com/kubernetes-engine/docs/concepts/alias-ips#cluster_sizing_primary_range Then we should choose /24 since the smallest possible subnet range is /24 when the secondary range assignment method is user-managed: https://cloud.google.com/kubernetes-engine/docs/concepts/alias-ips#cluster_sizing_secondary_range_pods Finally, Since we can use the whole range of secondary IP addresses, /22 will be enough as it provides 1024 IP addresses for our services. https://cloud.google.com/kubernetes-engine/docs/concepts/alias-ips#cluster_sizing_secondary_range_svcs
Ooh. Dat question, really make me confused. For this question not working " math and typical subnet". For example with typical math /28 mask provide 14 ip. But for primary configuration for node, only 12... After researching a lot of documents. I can write some notes: "Similarly, if you set the default maximum Pods to 8 and the cluster's secondary IP address range for Pods to /21, Kubernetes assigns a /28 CIDR range to nodes. This allows a maximum of 2(28-21) = 27 = 128 nodes on the cluster." source: https://cloud.google.com/kubernetes-engine/docs/how-to/flexible-pod-cidr#configuring_maximum_pods_per_node for secondary "There are no reserved IP addresses in the secondary IP ranges." source: https://cloud.google.com/vpc/docs/subnets#reserved_ip_addresses_in_every_subnet then decide for yourself which option is correct
A. Configure a /28 primary IP address range for the node IP addresses. This gives you 16 possible IP addresses, which is enough for your nodes. Configure a /25 secondary IP range for the Pods. This gives you 128 possible IP addresses, which is enough for your Pods. Configure a /22 secondary IP range for the Services. This gives you 1024 possible IP addresses, which is enough for your services.
Configure a /28 primary IP address range for the node IP addresses. This will allow up to 16 nodes in your cluster (14 usable plus 2 reserved). For example: 10.0.0.0/28 Configure a /25 secondary IP range for the Pods. This will allow up to 128 Pods per node (126 usable plus 2 reserved). For example: 10.0.1.0/25 Configure a /22 secondary IP range for the Services. This will allow up to 1024 Services in your cluster (1022 usable plus 2 reserved). For example: 10.0.4.0/22 Therefore, the correct answer is A.
21 is musi be so B
Sorry D as per: https://cloud.google.com/kubernetes-engine/docs/concepts/alias-ips#cluster_sizing_secondary_range_svcs
A , as /22 is 2^10-2 = 2046 services
I would tend to agree with Answer D. However it say 'Pads' versus 'Pods'. Is this a typo on the exam? If it is NOT a typo, then Answer D is wrong.