Misunderstanding in Lab 4.8

vadim.vararu

Hi all,

I'm not sure if there is a mistake in lab 4.8 or there is something i misunderstood.
It says to use the file design-review1.yaml and edit the pod resource requirements such that the CPU limit is exactly twice the amount requested by the container.(Hint: subtract .22)

Having the requested cpu of 0.3, to set the limit exactly twice the amount requested by the container means 0.6.
I don't understand what the hint subtract .22 means. Any ideas?

akovi

Well, this task is all over the place The stress is passed 1 cpu, it requests only 0.3 and limits to 2.22. But judged by the remark, we should suppose that the 0.3 should actually be 1.0. Opening a ticket may not be a bad idea, though to remove the ambiguity. (There are a few other instances of faults like this in the material)

fcioanca

Hi @akovi

All content-related questions or issues must be posted in the forum, and the course author and forum moderators will respond. Opening support tickets for content-related issues will only redirect you to the forum.

Thank you,
Flavia
The Linux Foundation Training Team

chrispokorni

Hi @vadim.vararu and @akovi,

There are three distinct properties and values for the CPU in the yaml definition manifest above.

limit: cpu=2.22 defines the ceiling value of a the CPU resource a container may be allowed to reach. Once exceeding the value the CPU consumption of the container may be throttled by Kubernetes.

request: cpu=0.3 defines the minimum value of the CPU resource that a container will need during init time. This is a value that is taken into account during scheduling time by the Kubernetes scheduler.

args: cpu=1 defines the actual amount of CPU that the container will attempt to consume.

From here, following the hint to "subtract .22", will yield the double (or "twice") the resource amount you are looking for.

No correction needed, just an understanding of the significance of the properties defined by the yaml definition manifest.

Regards,
-Chris

jacq0807

I struggled in this exercise as well, I was paying attention only on the limits and requests and without having kubectl top installed in my cp node, I am doing a trial-&-errror approach not until I tried looking at the mem-total and mem-alloc-size args. This will be a hint to anyone stuck in this section.