Installing N3000 daemon set RHCOS 4.8 fail

[dav1x]

It appears that as of 4.8 RHCOS has a read only /boot. The n3000 ds tries to create a sub-dir under /boot as part of the deployment process.

0s          Normal    Pulled                pod/n3000-daemonset-w45q8                           Container image "dhcp19-17-244.clus2.t5g.lab.eng.bos.redhat.com:5000/n3000-daemon:v1.1.0--1.3.8-2" already present on machine
0s          Warning   Failed                pod/n3000-daemonset-w45q8                           Error: failed to mkdir /boot/kernel: mkdir /boot/kernel: read-only file system
0s          Normal    Pulled                pod/n3000-daemonset-w45q8                           Container image "dhcp19-17-244.clus2.t5g.lab.eng.bos.redhat.com:5000/n3000-daemon:v1.1.0--1.3.8-2" already present on machine

More information regarding RHCOS.

coreos/fedora-coreos-config#356

I manually remounted /boot and the daemonset deployed properly

/usr/bin/mount -o remount,rw /boot

[root@rh8-tools n3k]# oc get pod
NAME                                           READY   STATUS    RESTARTS   AGE
accelerator-discovery-8ssxx                    1/1     Running   0          74m
fpga-driver-daemonset-h2psb                    1/1     Running   0          74m
fpgainfo-exporter-nlg87                        1/1     Running   0          74m
n3000-controller-manager-695d586f69-76624      2/2     Running   0          75m
n3000-daemonset-w45q8                          1/1     Running   0          74m
sriov-device-plugin-q99ft                      1/1     Running   0          74m
sriov-fec-controller-manager-6896c47b8-kgzpl   2/2     Running   0          75m
sriov-fec-daemonset-8jp8s                      1/1     Running   0          74m
[root@rh8-tools n3k]# oc get n3000nodes.fpga.intel.com 
NAME                                      FLASH
cnfdb4.clus2.t5g.lab.eng.bos.redhat.com   NotRequested

[rpieczon]

This repository doesn't contain n3k operator anymore and n3000 operator is EOL now.