Lab 4.3 issues

risicle

There appear to be a few things wrong with Lab 4.3 "Resource Limits for a Namespace"

Firstly, the new deployment created in step 5 is never given any arguments, so will never really occupy any resources.

Secondly, the instructions seem to go to pains to limit the memory of a hog to 500M, while seemingly asking it to attempt to allocate 950M. Then it has a top listing at the end (step 12) showing two processes happily using ~950M each. Are we lead to conclude that the memory limit has no effect?

If the instructions are followed verbatim, with the vish/stress image available at time of writing, what you end up with is the first deployment happily continuing to use 950M and the second deployment in a crash-loop presumably failing to allocate the memory it was asked to.

This is all quite confusing.

chrispokorni

Hi @risicle,

The first deployment in lab 4.3 should exemplify the pod inheriting the namespace cpu and memory limits, as defined by the limitrange resource, while the hog2.yaml definition manifest allows for the args to be modified for the second deployment, if required.

Regards,
-Chris

risicle

Thanks Chris, so it's relying on inheriting the args added to hog.yaml in the previous section. Guess that makes sense in a way.

I still don't really understand about the 950M/500M memory limit incongruity though.