Lab4.3 content inaccurancy

lioneyes · April 12

Step 12 at Exercise 4.3: "Resource Limits for a Namespace" describes a top command output that is equal for both default and low-usage-limit namespace. But hog deployment prepared in lab 4.2 is configured to consume 2 cpus. So in limited namespace stress container really consumes 100 cpu (in top) as it limited by namespace. But default namespaced hog container actually consumes 200 cpu (in top) as expected, isn't it?

chrispokorni · April 21

Hi @lioneyes,

The hog deployment from lab exercise 4.2 is allowed to consume close to 100% CPU, per the CPU limit (1) defined in the manifest. This reading is consistent with a node/VM with 2 CPU cores.

The hog deployment from lab exercise 4.3 is restricted by the limitrange CPU limit (1) set on the newly created namespace, therefore the reading is also close to 100%. This reading is also consistent with a node/VM with 2 CPU cores.

Are your top readings different?

Regards,
-Chris

lioneyes · April 30

I forgot to add "cpu: 1" into hog's manifest limits map...
So it consumes 200% cpu as requested in args section for unlimited namespace and it consumes 100% cpu in limited one as expected with such a limit range policy.

So it was my mistake and a good lesson about how limits work)

p.s. my manifest for better understanding

     resources:
      limits:
        memory: "4Gi"
      requests:
        memory: "2500Mi"
    args:
    - -cpus
    - "2"
    - -mem-total
    - "950Mi"
    - -mem-alloc-size
    - "100Mi"
    - -mem-alloc-sleep
    - "1s"

Lab4.3 content inaccurancy

Answers

Welcome!

Welcome!

Quick Links

Categories

Upcoming Training

Kubernetes Administration (LFS458)

Linux System Administration (LFS301)

Open Source Virtualization (LFS462)

Linux Kernel Debugging and Security (LFD440)