Bug 2002313 - Kubevirt doesn't respect cpuAllocationRatio
Summary: Kubevirt doesn't respect cpuAllocationRatio
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Container Native Virtualization (CNV)
Classification: Red Hat
Component: Virtualization
Version: 4.8.1
Hardware: Unspecified
OS: Unspecified
high
high
Target Milestone: ---
: 4.9.1
Assignee: Vladik Romanovsky
QA Contact: Kedar Bidarkar
URL:
Whiteboard:
Depends On:
Blocks: 2004950
TreeView+ depends on / blocked
 
Reported: 2021-09-08 14:17 UTC by Erkan Erol
Modified: 2021-12-13 19:59 UTC (History)
4 users (show)

Fixed In Version: virt-operator-container-v4.9.1-4 hco-bundle-registry-container-v4.9.1-15
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
: 2004950 (view as bug list)
Environment:
Last Closed: 2021-12-13 19:59:01 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Github kubevirt kubevirt pull 6423 0 None open [release-0.44] Fix the default setting of CPU requests on vmipods 2021-11-02 13:28:12 UTC
Red Hat Product Errata RHBA-2021:5091 0 None None None 2021-12-13 19:59:17 UTC

Description Erkan Erol 2021-09-08 14:17:40 UTC
Description of problem:


I installed OpenShift Virtualization 4.8.1 and created different VMs with different specs. I used different numbers for cpu core field. However, Kubevirt doesn't change the requested cpu amount in virt-launcher spec. It was always 100m. 

Version-Release number of selected component (if applicable): 4.8.1


Actual results:
All virt-launcher pods contains the same amount of cpu. 100m


Expected results: It should be proportional to VM spec as explained here: http://kubevirt.io/user-guide/operations/node_overcommit/#node-cpu-allocation-ratio


Additional info: HCO is not setting cpuAllocationRatio in Kubevirt and it is empty by default.

Comment 1 Kedar Bidarkar 2021-09-08 18:45:46 UTC
Tried to reproduce this:

a) Explicitly set the "cpuAllocationRatio" value via the JSON Annotation Patch.

]$ oc annotate --overwrite -n openshift-cnv hyperconverged kubevirt-hyperconverged kubevirt.kubevirt.io/jsonpatch='[{
      "op": "add",
      "path": "/spec/configuration/developerConfiguration/cpuAllocationRatio",
      "value": 20
  }]'

b) Value appears set correctly

]$ oc get kubevirt kubevirt-kubevirt-hyperconverged -n openshift-cnv -o yaml | grep -A 5 "configuration:"
  configuration:
    developerConfiguration:
      cpuAllocationRatio: 20
      diskVerification:
        memoryLimit: 2G
      featureGates:

c) I had set the "cores" count under "cpu" as 6
]$ oc get vmi vm22-rhel84-ocs -o yaml | grep -A 5 "cpu:"
    cpu:
      cores: 6
      sockets: 1
      threads: 1
---

d) virt-launcher pod shows "cpu: 100m" 

]$ oc get pod virt-launcher-vm22-rhel84-ocs-sv46l -o yaml | grep "cpu:"
        cpu: 100m

POD CPU request = number of vCPUs * 1/cpuAllocationRatio

Probably the expected value should have been  6/20 = 330m or 300m , IIUC

---
e) 
Logged into the VM and checked for the nproc value. It is indeed 6.

[cloud-user@vm22-rhel84-ocs ~]$ nproc
6

Comment 2 Kedar Bidarkar 2021-09-08 18:47:41 UTC
comment1 stuff was reproduced on CNV-4.9.0.

Comment 3 Vladik Romanovsky 2021-09-13 15:30:09 UTC
I've posted https://github.com/kubevirt/kubevirt/pull/6374 to fix this.

Comment 4 Kedar Bidarkar 2021-11-09 12:54:46 UTC
With the following:
[kbidarka@localhost ocs]$ oc get vmi vm-rhel84-ocs -o yaml | grep -A 5 "cpu:"
--
    cpu:
      cores: 6
      sockets: 1
      threads: 1

---


[kbidarka@localhost ocs]$ oc get pod virt-launcher-vm-rhel84-ocs-xkphg -o yaml | grep "cpu:"
                f:cpu: {}
        cpu: 300m
[kbidarka@localhost ocs]$ oc get kubevirt kubevirt-kubevirt-hyperconverged -n openshift-cnv -o yaml | grep -A 5 "configuration:"
--
  configuration:
    developerConfiguration:
      cpuAllocationRatio: 20
      diskVerification:
        memoryLimit: 2G
      featureGates:


---

With default value of "cpuAllocationRatio" as 10

[kbidarka@localhost ocs]$ oc get pod virt-launcher-vm-rhel84-ocs-g9kx7 -o yaml | grep "cpu:"
                f:cpu: {}
        cpu: 600m

---

VERIFIED with virt-operator-container-v4.9.1-4

Comment 10 errata-xmlrpc 2021-12-13 19:59:01 UTC
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory (OpenShift Virtualization 4.9.1 Images), and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHBA-2021:5091


Note You need to log in before you can comment on or make changes to this bug.