I Think Hight CPU Ready Times Are Caused by my Resource Pools - Need Some Help Understanding!

Posted by Photo-Josh@reddit | sysadmin | View on Reddit | 5 comments

Hey all, Picked up a new envornment in the last 6 months and trying to understand why we have a fair amount of CPU resource avaiable, but still high CPU ready times in our Preprod resource pool. I THINK it's due to how our resource pools are configured but not 100% sure. We have 2 resource pools: **Prod** and **Preprod**. Screenshots on how they're setup: https://imgur.com/a/k62Po8S What I **THINK** is going on: Although we have "enough" CPU resources within the cluster, because preprod is set to low, it'll never be able to contend with prod. However, if I set Prod to High, and Preprod to Normal.. I'll still have a 1:2 ratio of 4000 to 8000.. So I'm not sure if that'll solve my problem? I'm also not sure if I should set a CPU reservation / limit closer to what the actual usage is?

Reply to Post

5 Comments

[-]

gdelia928@reddit

I don’t think resource pools are your problem, at least not based on your description of the issue and the pics shared. From the graphs your showing I don’t see you approaching the limits your set on the cpu side for either pool, so assuming your cpu capacity is higher or close to the combined number resource pools are never doing anything as they only enforce shares when resource contention is high (e.g. your vms are using nearly all the copy the cluster has available). This is a common misconception about resource pools, they only jump into play when you run out of capacity to service your vms. More likely you have an issue with over allocation of vCPU to your individual vms and it doesn’t align well with your pCPU layout. What’s your vCPU:pCPU ratio? It’s also possible if you are setting limits on your vms to artificially cause high ready, but based on the graphs it wouldn’t be the resource pool limit causing the issue since it’s never approached, but individual vm cpu limits.

[-]

h0rnman@reddit

I'll agree on this one here. It doesn't look like resource pool issues. Do you have some wacky configuration where there is, for example, 40 cores on a host and you have 40 vcpu assigned to 3 or 4 of the vms (along with the normal partial allocation)? That can cause cpu ready issues even if you have plenty of resources available

[-]

HilkoVMware@reddit

Shares might not be your problem, more likely the limit is.

[-]

Photo-Josh@reddit (OP)

I never see it go over 80-90k CPU though for the Preprod pool? Did you see the graph in the photo album I linked? Unless I'm reading that wrong?

[-]

HilkoVMware@reddit

Not really, it’s blurry and says I need to install an app. But limits are hard, and shares only come into play when there is contention.