Bug 1912507

Summary: unable to use hugepages
Product: OpenShift Container Platform Reporter: Qian Cai <qcai>
Component: sandboxed-containersAssignee: Ariel Adam <aadam>
Status: CLOSED DEFERRED QA Contact: Cameron Meadors <cmeadors>
Severity: high Docs Contact:
Priority: high    
Version: 4.8CC: aadam, gkurz, wlewis
Target Milestone: ---   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2023-03-09 01:00:39 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:

Description Qian Cai 2021-01-04 15:37:51 UTC 
Description of problem:
This uses to track to hugepages feature for sandboxed containers including the upstream merging (kata 2.0) and then backport to the downstream.

Following the example in https://docs.openshift.com/container-platform/4.5/scalability_and_performance/what-huge-pages-do-and-how-they-are-consumed-by-apps.html

results in creating pod failures.

Version-Release number of selected component (if applicable):
OCP 4.8

How reproducible:
always

Steps to Reproduce:
1) make sure the node has 100M hugepages preallocated.

$ oc describe nodes | grep huge

hugepages-2Mi: 100Mi

$ cat <<EOF> deploy/huge.yaml

apiVersion: v1
kind: Pod
metadata:
  generateName: hugepages-volume-
spec:
  containers:
  - securityContext:
      privileged: true
    image: rhel7:latest
    command:
    - sleep
    - inf
    name: example
    volumeMounts:
    - mountPath: /dev/hugepages
      name: hugepage
    resources:
      limits:
        hugepages-2Mi: 100Mi 
        memory: "1Gi"
        cpu: "1"
  volumes:
  - name: hugepage
    emptyDir:
      medium: HugePages
  runtimeClassName: kata-oc

EOF

$ oc create -f huge.yaml

$ oc describe pod hugepages-volume-4x5nn
...
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Scheduled 115s default-scheduler Successfully assigned default/hugepages-volume-4x5nn to kataqe-nmfh9-worker-rv2nh
Normal AddedInterface 108s multus Add eth0 [14.128.2.19/23]
Normal Pulling 13s (x6 over 99s) kubelet, kataqe-nmfh9-worker-rv2nh Pulling image "rhel7:latest"
Normal Pulled 11s (x6 over 89s) kubelet, kataqe-nmfh9-worker-rv2nh Successfully pulled image "rhel7:latest"
Warning Failed 8s (x6 over 84s) kubelet, kataqe-nmfh9-worker-rv2nh Error: CreateContainer failed: Timeout reached after 3s waiting for device 0:0:0:0/block: unknown
Comment 1 Qian Cai 2021-01-04 20:22:17 UTC 
For a record, Pradipta mentioned there are out-of-tree repos for kata 1.x only.

https://github.com/bpradipt/agent/tree/hugepages
https://github.com/bpradipt/runtime/tree/hugepages
Comment 7 Shiftzilla 2023-03-09 01:00:39 UTC 
OpenShift has moved to Jira for its defect tracking! This bug can now be found in the OCPBUGS project in Jira.

https://issues.redhat.com/browse/OCPBUGS-8829