Closed
Description
Hi guys, I am trying Intel FPGA Device PlugIn with A10 FPGA.
The FPGA resource has report correctly, I checked it by kubectl describe nodes 127.0.0.1 | grep -A4 Alloc, the output is :
Allocatable:
cpu: 32
ephemeral-storage: 3188670572180
fpga.intel.com/af-f7df405cbd7acf7222f144b0b93acd18: 1
hugepages-1Gi: 4Gi
I have succeed run a pod with this yaml file:
apiVersion: v1
kind: Pod
metadata:
name: test-nlb3
spec:
containers:
- name: test-nlb3
image: intel/opae-nlb-demo:devel
imagePullPolicy: IfNotPresent
command: ['nlb3']
securityContext:
capabilities:
add:
[IPC_LOCK]
resources:
limits:
fpga.intel.com/af-f7df405cbd7acf7222f144b0b93acd18: 1
cpu: 1
hugepages-2Mi: 20Mi
restartPolicy: Never
But if I changed the resources section to:
resources:
limits:
fpga.intel.com/arria10.dcp1.0-nlb3: 1
cpu: 1
hugepages-2Mi: 20Mi
The pod will be pending with an schedule error.
The related logs are the following:
Run kubectl logs intel-fpga-webhook-deployment-7f6b8474c-nd64g
2020/04/29 12:15:55 http: TLS handshake error from 172.17.0.1:23786: remote error: tls: bad certificate
Run tail -f kube-apiserver.log:
failed calling webhook "fpga.mutator.webhooks.intel.com": Post "https://intel-fpga-webhook-svc.default.svc:443/pods?timeout=30s": x509: certificate signed by unknown authority (possibly because of "crypto/rsa: verification error" while trying to verify candidate authority certificate "ca")
Run kubectl describe pod test-nlb3:
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning FailedScheduling 6s default-scheduler 0/1 nodes are available: 1 Insufficient fpga.intel.com/arria10.dcp1.0-nlb3.
Warning FailedScheduling 6s default-scheduler 0/1 nodes are available: 1 Insufficient fpga.intel.com/arria10.dcp1.0-nlb3.
Also, I wonder the notion of DCP, I noticed there are 3 version of DCPs, I have tried each of them, and had always the same error. What does DCP exactly mean?
Metadata
Metadata
Assignees
Labels
No labels