116 lines
4.0 KiB
PowerShell
116 lines
4.0 KiB
PowerShell
param(
|
|
[Parameter(Mandatory = $true)]
|
|
[ValidateSet("vllm-text", "llama-gguf", "translation-api", "audio-api", "video-api", "image-api", "model-chat-ui")]
|
|
[string]$Service,
|
|
|
|
[Parameter(Mandatory = $true)]
|
|
[ValidateSet("start", "stop", "restart", "status", "logs")]
|
|
[string]$Action,
|
|
|
|
[string]$DockerHost = "ssh://docker-gpu",
|
|
[string]$SshTarget = "docker-gpu",
|
|
[int]$Tail = 80
|
|
)
|
|
|
|
$ErrorActionPreference = "Stop"
|
|
|
|
$services = @{
|
|
"vllm-text" = @{
|
|
Compose = "core/deploy/docker-gpu/vllm/compose.yaml"
|
|
Env = "core/deploy/docker-gpu/vllm/.env.windows-z.example"
|
|
Project = "llm-vllm-text"
|
|
Container = "llm-vllm-text"
|
|
}
|
|
"llama-gguf" = @{
|
|
Compose = "core/deploy/docker-gpu/llama-cpp/compose.yaml"
|
|
Env = "core/deploy/docker-gpu/llama-cpp/.env.windows-z.example"
|
|
Project = "llm-llama-gguf"
|
|
Container = "llm-llama-devstral-1c"
|
|
ExtraContainers = @("llm-llama-qwen3-coder-q6-test")
|
|
}
|
|
"translation-api" = @{
|
|
Compose = "core/deploy/docker-gpu/transformers/translation.compose.yaml"
|
|
Env = "core/deploy/docker-gpu/transformers/translation.env.example"
|
|
Project = "llm-transformers-translation"
|
|
Container = "llm-transformers-translation"
|
|
}
|
|
"audio-api" = @{
|
|
Compose = "core/deploy/docker-gpu/transformers/audio.compose.yaml"
|
|
Env = "core/deploy/docker-gpu/transformers/audio.env.example"
|
|
Project = "llm-transformers-audio"
|
|
Container = "llm-transformers-audio"
|
|
}
|
|
"video-api" = @{
|
|
Compose = "core/deploy/docker-gpu/transformers/video.compose.yaml"
|
|
Env = "core/deploy/docker-gpu/transformers/video.env.example"
|
|
Project = "llm-transformers-video"
|
|
Container = "llm-transformers-video"
|
|
}
|
|
"image-api" = @{
|
|
Compose = "core/deploy/docker-gpu/transformers/image.compose.yaml"
|
|
Env = "core/deploy/docker-gpu/transformers/image.env.example"
|
|
Project = "llm-transformers-image"
|
|
Container = "llm-transformers-image"
|
|
}
|
|
"model-chat-ui" = @{
|
|
Compose = "core/deploy/docker-gpu/model-chat/compose.yaml"
|
|
Env = "core/deploy/docker-gpu/model-chat/.env.windows-z.example"
|
|
Project = "llm-model-chat-ui"
|
|
Container = "llm-model-chat-ui"
|
|
}
|
|
}
|
|
|
|
$plan = $services[$Service]
|
|
if (-not (Test-Path -LiteralPath $plan.Compose)) {
|
|
throw "Compose file not found: $($plan.Compose)"
|
|
}
|
|
|
|
$composeArgs = @("--host", $DockerHost, "compose", "-p", $plan.Project)
|
|
if ($plan.Env -and (Test-Path -LiteralPath $plan.Env)) {
|
|
$composeArgs += @("--env-file", $plan.Env)
|
|
}
|
|
$composeArgs += @("-f", $plan.Compose)
|
|
|
|
switch ($Action) {
|
|
"start" {
|
|
$existing = docker --host $DockerHost ps -a --filter "name=^$($plan.Container)$" --format "{{.Names}}"
|
|
if ($existing -contains $plan.Container) {
|
|
docker --host $DockerHost start $plan.Container
|
|
} else {
|
|
docker @composeArgs up -d
|
|
}
|
|
}
|
|
"stop" {
|
|
$containers = @($plan.Container)
|
|
if ($plan.ExtraContainers) {
|
|
$containers += $plan.ExtraContainers
|
|
}
|
|
foreach ($container in $containers) {
|
|
$existing = docker --host $DockerHost ps -a --filter "name=^$container$" --format "{{.Names}}"
|
|
if ($existing -contains $container) {
|
|
docker --host $DockerHost stop $container
|
|
}
|
|
}
|
|
}
|
|
"restart" {
|
|
docker --host $DockerHost restart $plan.Container
|
|
}
|
|
"status" {
|
|
$containers = @($plan.Container)
|
|
if ($plan.ExtraContainers) {
|
|
$containers += $plan.ExtraContainers
|
|
}
|
|
foreach ($container in $containers) {
|
|
docker --host $DockerHost ps -a --filter "name=^$container$"
|
|
}
|
|
Write-Host ""
|
|
Write-Host "GPU:"
|
|
ssh $SshTarget "nvidia-smi --query-gpu=index,name,memory.used,memory.total,utilization.gpu --format=csv,noheader,nounits"
|
|
}
|
|
"logs" {
|
|
docker --host $DockerHost logs --tail $Tail $plan.Container
|
|
}
|
|
}
|
|
|
|
exit $LASTEXITCODE
|