mirror of
https://github.com/cozystack/cozystack.git
synced 2026-10-29 08:07:33 +09:00
Aleksei Sviridkin
dca23da220
GPUNode is the hoster's per-node view of GPUs and the binder's source of device health, so something on each node has to report it. The agent reads the GPU functions from sysfs and what kubelet's device plugins make of them from the pod-resources API. Plugins report devices by PCI address, mdev UUID or IOMMU group, so each ID is resolved through sysfs by its shape rather than by knowing the plugin. A group counts only through functions bound to vfio, so the plain indexes NVIDIA's container plugin may report do not read as groups. GPUs of NVIDIA's container plugin, reported by NVML UUID, cannot be placed through sysfs, but their resource name still joins resourceNames, since pods request GPUs by it. pod-resources carries no health; kubelet lists only healthy devices as allocatable, which is what the agent reads health from, and it counts them per function, because an mdev parent serves several. Pods in a final phase are skipped because kubelet 1.32 and 1.33 still list them with devices a new pod may hold. The binder writes conditions into the same status, so the agent only merge-patches its own fields, and it writes no GPUNode but its own node's. Assisted-by: LLM Signed-off-by: Aleksei Sviridkin <f@lex.la>
54 lines
1.5 KiB
Go
54 lines
1.5 KiB
Go
/*
|
|
Copyright 2026 The Cozystack Authors.
|
|
|
|
Licensed under the Apache License, Version 2.0 (the "License");
|
|
you may not use this file except in compliance with the License.
|
|
You may obtain a copy of the License at
|
|
|
|
http://www.apache.org/licenses/LICENSE-2.0
|
|
|
|
Unless required by applicable law or agreed to in writing, software
|
|
distributed under the License is distributed on an "AS IS" BASIS,
|
|
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
See the License for the specific language governing permissions and
|
|
limitations under the License.
|
|
*/
|
|
|
|
package agent
|
|
|
|
// models names the NVIDIA device IDs of the platform's passthrough defaults
|
|
// (packages/core/platform/files/gpu-passthrough-defaults.yaml); a test keeps
|
|
// the two in step. The host carries no PCI ID database to read them from.
|
|
var models = map[string]string{
|
|
"2335": "H200 SXM 141GB",
|
|
"2330": "H100 SXM5 80GB",
|
|
"2331": "H100 PCIe",
|
|
"2321": "H100L 94GB",
|
|
"26b9": "L40S",
|
|
"26b5": "L40",
|
|
"27b8": "L4",
|
|
"20b2": "A100 SXM4 80GB",
|
|
"20b0": "A100 SXM4 40GB",
|
|
"20b5": "A100 PCIe 80GB",
|
|
"20f1": "A100 PCIe 40GB",
|
|
"20b7": "A30 PCIe",
|
|
"2235": "A40",
|
|
"2236": "A10",
|
|
"1eb8": "T4",
|
|
"1db1": "V100 SXM2 16GB",
|
|
"1db5": "V100 SXM2 32GB",
|
|
"1db4": "V100 PCIe 16GB",
|
|
"1df6": "V100S PCIe 32GB",
|
|
}
|
|
|
|
// model falls back to <vendor>-<device>, the form of the per-model node
|
|
// labels.
|
|
func model(vendor, device string) string {
|
|
if vendor == vendorNVIDIA {
|
|
if m, ok := models[device]; ok {
|
|
return m
|
|
}
|
|
}
|
|
return vendor + "-" + device
|
|
}
|