[BUG] the engine image controller has to deploy the engine images with the linux node selector
doctorpangloss opened this issue · comments
Describe the bug
the engine image controller tries to deploy images onto windows nodes, which can't run them
the daemonset (along with the csi-attacher, csi-provisioner, csi-snapshotter and csi-resizer deployments) need to have the proper node selectors for linux instead of windows
To Reproduce
- deploy longhorn onto a windows & linux cluster
- observe failures to deploy the engine image on windows nodes, which shouldn't be scheduled in the first place
Expected behavior
these resources need to have configurable node selectors, affinity, tolerations, etc.
Support bundle for troubleshooting
N/A
Environment
- Longhorn version: v1.6.2
- Installation method: flux
- Kubernetes distro: k02
- Number of control plane nodes in the cluster: 3
- Number of worker nodes in the cluster: 20
- Node config
- OS type and version: Windows 2022, Ubuntu 22
- Kernel version: 6.5.0-35-generic
- CPU per node: ~16-64 Zen4 class
- Memory per node: 32-128GB typical
- Disk type (e.g. SSD/NVMe/HDD): NVMe
- Network bandwidth between the nodes (Gbps): 2.5gbps-10gbps typical
- Underlying Infrastructure (e.g. on AWS/GCE, EKS/GKE, VMWare/KVM, Baremetal): Baremetal
- Number of Longhorn volumes in the cluster: 40
cc @c3y1huang