Skip to content

Instantly share code, notes, and snippets.

@c4software

c4software/setup-llm.sh

Created June 17, 2026 20:19
Show Gist options

  • Select an option

    Learn more about clone URLs
  • Save c4software/9c2707dfa1b5f3f650a5c75498aae741 to your computer and use it in GitHub Desktop.

Select an option

Learn more about clone URLs
Save c4software/9c2707dfa1b5f3f650a5c75498aae741 to your computer and use it in GitHub Desktop.
Download ZIP
llama server multi models setup (Simple script to setup LLM Server for Ryzen Strix Halo 128Gb)
Raw
setup-llm.sh
#!/usr/bin/env bash
set -euo pipefail
# =============================================================================
# setup-llm.sh — llama-server router mode natif
# =============================================================================
# =============================================================================
# MODÈLES (repo + fichier)
# =============================================================================
MODEL_QWEN35_2B_REPO="unsloth/Qwen3.5-2B-GGUF"
MODEL_QWEN35_2B_FILE="Qwen3.5-2B-UD-Q4_K_XL.gguf"
MODEL_QWEN35_9B_REPO="unsloth/Qwen3.5-9B-GGUF"
MODEL_QWEN35_9B_FILE="Qwen3.5-9B-UD-Q6_K_XL.gguf"
MODEL_QWEN36_35B_A3B_REPO="unsloth/Qwen3.6-35B-A3B-GGUF"
MODEL_QWEN36_35B_A3B_FILE="Qwen3.6-35B-A3B-UD-Q6_K_XL.gguf"
MODEL_QWEN3_CODER_NEXT_REPO="unsloth/Qwen3-Coder-Next-GGUF"
MODEL_QWEN3_CODER_NEXT_FILE="Qwen3-Coder-Next-UD-Q4_K_XL.gguf"
MODEL_QWEN36_27B_REPO="unsloth/Qwen3.6-27B-GGUF"
MODEL_QWEN36_27B_FILE="Qwen3.6-27B-UD-Q6_K_XL.gguf"
# Gemma 4 31B — MTP drafter embarqué dans le repo principal
MODEL_GEMMA_31B_REPO="unsloth/gemma-4-31B-it-GGUF"
MODEL_GEMMA_31B_FILE="gemma-4-31B-it-Q4_K_M.gguf"
MODEL_GEMMA_31B_MTP_FILE="mtp-gemma-4-31B-it.gguf"
# Gemma 4 12B — modèle unifié texte+image+audio+vidéo, MTP drafter embarqué
MODEL_GEMMA_12B_REPO="unsloth/gemma-4-12b-it-GGUF"
MODEL_GEMMA_12B_FILE="gemma-4-12b-it-UD-Q4_K_XL.gguf"
MODEL_GEMMA_12B_MTP_FILE="mtp-gemma-4-12b-it.gguf"
# GPT-OSS 120B — shards UD-Q4_K_XL
MODEL_GPTOSS_REPO="unsloth/gpt-oss-120b-GGUF"
MODEL_GPTOSS_FILE_GLOB="UD-Q4_K_XL/*"
MODEL_GPTOSS_FILE_ENTRY="UD-Q4_K_XL/gpt-oss-120b-UD-Q4_K_XL-00001-of-00002.gguf"
# Qwen3.6-27B-MTP — variante MTP, draft intégré (Q4_K_XL)
MODEL_QWEN36_27B_MTP_REPO="unsloth/Qwen3.6-27B-MTP-GGUF"
MODEL_QWEN36_27B_MTP_FILE="Qwen3.6-27B-UD-Q4_K_XL.gguf"
# Qwen3.6-35B-A3B-MTP — variante MTP, draft intégré (Q4_K_XL)
MODEL_QWEN36_35B_A3B_MTP_REPO="unsloth/Qwen3.6-35B-A3B-MTP-GGUF"
MODEL_QWEN36_35B_A3B_MTP_FILE="Qwen3.6-35B-A3B-UD-Q4_K_XL.gguf"
# MiniMax-M2.7 — MoE 230B (10B actifs), 200K ctx, shards UD-IQ4_XS
MODEL_MINIMAX_REPO="unsloth/MiniMax-M2.7-GGUF"
MODEL_MINIMAX_FILE_GLOB="UD-IQ4_XS/*"
MODEL_MINIMAX_FILE_ENTRY="UD-IQ4_XS/MiniMax-M2.7-UD-IQ4_XS-00001-of-00004.gguf"
# Qwopus3.6-27B-Coder-MTP — fine-tune coder SFT, SWE-bench Verified 67.0%
MODEL_QWOPUS_CODER_MTP_REPO="Jackrong/Qwopus3.6-27B-Coder-MTP-GGUF"
MODEL_QWOPUS_CODER_MTP_FILE="Qwopus3.6-27B-Coder-MTP-Q5_K_M.gguf"
# =============================================================================
# CHEMINS
# =============================================================================
MODELS_BASE="$HOME/models"
CONFIG_DIR="$HOME/models"
QWEN35_2B_PATH="$MODELS_BASE/qwen3.5-2b/$MODEL_QWEN35_2B_FILE"
QWEN35_9B_PATH="$MODELS_BASE/qwen3.5-9b/$MODEL_QWEN35_9B_FILE"
QWEN36_35B_A3B_PATH="$MODELS_BASE/qwen3.6-35b-a3b/$MODEL_QWEN36_35B_A3B_FILE"
QWEN3_CODER_NEXT_PATH="$MODELS_BASE/qwen3-coder-next/$MODEL_QWEN3_CODER_NEXT_FILE"
QWEN36_27B_PATH="$MODELS_BASE/qwen3.6-27b/$MODEL_QWEN36_27B_FILE"
GEMMA_31B_PATH="$MODELS_BASE/gemma-31b/$MODEL_GEMMA_31B_FILE"
GEMMA_31B_MTP_PATH="$MODELS_BASE/gemma-31b/$MODEL_GEMMA_31B_MTP_FILE"
GEMMA_12B_PATH="$MODELS_BASE/gemma-12b/$MODEL_GEMMA_12B_FILE"
GEMMA_12B_MTP_PATH="$MODELS_BASE/gemma-12b/$MODEL_GEMMA_12B_MTP_FILE"
GPTOSS_PATH="$MODELS_BASE/gpt-oss/$MODEL_GPTOSS_FILE_ENTRY"
QWEN36_27B_MTP_PATH="$MODELS_BASE/qwen3.6-27b-mtp/$MODEL_QWEN36_27B_MTP_FILE"
QWEN36_35B_A3B_MTP_PATH="$MODELS_BASE/qwen3.6-35b-a3b-mtp/$MODEL_QWEN36_35B_A3B_MTP_FILE"
MINIMAX_PATH="$MODELS_BASE/minimax/$MODEL_MINIMAX_FILE_ENTRY"
QWOPUS_CODER_MTP_PATH="$MODELS_BASE/qwopus3.6-27b-coder-mtp/$MODEL_QWOPUS_CODER_MTP_FILE"
# =============================================================================
# PRESETS INI
#
# Chaque entrée du tableau associatif définit le corps d'un preset [section].
# Conventions :
# - Une clé ini par ligne, format "key = value"
# - Les flags globaux ([*]) ne sont pas répétés ici
# - Commentaire de section : ligne(s) commençant par ";"
# - load-on-startup et stop-timeout gèrent le cycle de vie (always-on vs LRU)
# - cache-reuse = 0 écrase le global 4096 pour les modèles MoE/MTP incompatibles
# - cache-type-v = q8_0 en surcharge locale pour les coding agents
# - swa-full + ctx-checkpoints : hybrides Qwen3.5/3.6 uniquement
# - spec-draft-model : drafter externe (Gemma MTP uniquement)
# =============================================================================
declare -A MODEL_INI
# Qwen3.5-2B : dense 2B hybrid-thinking, ultra-léger
# swa-full : exploite le ctx complet (SWA hybride GatedDeltaNet)
MODEL_INI[qwen3.5-2b]="
model = $QWEN35_2B_PATH
ctx-size = 32768
cache-ram = 2048
temp = 0.7
top-k = 20
top-p = 0.8
min-p = 0.0
parallel = 4
swa-full = true
ctx-checkpoints = 128"
# Qwen3.5-9B : dense 9B hybrid-thinking
MODEL_INI[qwen3.5-9b]="
model = $QWEN35_9B_PATH
ctx-size = 32768
cache-ram = 2048
temp = 0.7
top-k = 20
top-p = 0.8
min-p = 0.0
parallel = 4
swa-full = true
ctx-checkpoints = 128
load-on-startup = true"
# Qwen3.6-35B-A3B nothink — MoE hybride, always-on
# cache-reuse 0 : MoE hybride incompatible avec le cache-reuse global
MODEL_INI[qwen3.6-35b-a3b-nothink]="
model = $QWEN36_35B_A3B_PATH
ctx-size = 524288
cache-ram = 12288
reasoning = off
temp = 0.6
top-k = 20
top-p = 0.95
min-p = 0.0
parallel = 2
cache-reuse = 0
swa-full = true
ctx-checkpoints = 128
load-on-startup = true"
# Qwen3.6-35B-A3B thinking — raisonnement activé
# cache-reuse 0 : MoE hybride incompatible
MODEL_INI[qwen3.6-35b-a3b]="
model = $QWEN36_35B_A3B_PATH
ctx-size = 393216
cache-ram = 6144
temp = 0.6
top-k = 20
top-p = 0.95
min-p = 0.0
parallel = 3
cache-reuse = 0
swa-full = true
ctx-checkpoints = 128
stop-timeout = 300"
# Qwen3.6-35B-A3B-MTP nothink — spéculation MTP draft=2
# cache-reuse 0 : incompatible MTP — nothink via chat-template-kwargs (pas --reasoning off)
MODEL_INI[qwen3.6-35b-a3b-mtp-nothink]="
model = $QWEN36_35B_A3B_MTP_PATH
ctx-size = 131072
cache-ram = 6144
temp = 0.7
top-k = 20
top-p = 0.8
min-p = 0.0
presence-penalty = 1.5
chat-template-kwargs = {\"enable_thinking\":false}
cache-reuse = 0
spec-type = draft-mtp
spec-draft-n-max = 2
jinja = true
parallel = 1
swa-full = true
ctx-checkpoints = 128
stop-timeout = 600"
# Qwen3-Coder-Next — MoE 80B hybrid-attention, agentic coding
# cache-type-v q8_0 : précision V critique pour les diffs de code
# cache-reuse 0 : MoE hybrid-attention incompatible
MODEL_INI[qwen3-coder-next]="
model = $QWEN3_CODER_NEXT_PATH
ctx-size = 131072
cache-ram = 4096
temp = 1.0
top-k = 40
top-p = 0.95
min-p = 0.01
cache-type-v = q8_0
cache-reuse = 0
parallel = 1
stop-timeout = 1800"
# Qwen3.6-27B thinking — raisonnement activé, tool-calling jinja, presence_penalty
MODEL_INI[qwen3.6-27b]="
model = $QWEN36_27B_PATH
ctx-size = 131072
cache-ram = 4096
temp = 0.6
top-k = 20
top-p = 0.95
min-p = 0.0
presence-penalty = 1.5
jinja = true
parallel = 1
swa-full = true
ctx-checkpoints = 128
stop-timeout = 600"
# Qwen3.6-27B nothink — réponses directes sans chain-of-thought
MODEL_INI[qwen3.6-27b-nothink]="
model = $QWEN36_27B_PATH
ctx-size = 131072
cache-ram = 4096
reasoning = off
temp = 0.6
top-k = 20
top-p = 0.95
min-p = 0.0
presence-penalty = 1.5
parallel = 1
swa-full = true
ctx-checkpoints = 128
stop-timeout = 600"
# Qwen3.6-27B-MTP nothink — spéculation MTP draft=2
# cache-reuse 0 : incompatible MTP — nothink via chat-template-kwargs
MODEL_INI[qwen3.6-27b-mtp-nothink]="
model = $QWEN36_27B_MTP_PATH
ctx-size = 131072
cache-ram = 4096
temp = 0.7
top-k = 20
top-p = 0.8
min-p = 0.0
presence-penalty = 1.5
chat-template-kwargs = {\"enable_thinking\":false}
cache-reuse = 0
spec-type = draft-mtp
spec-draft-n-max = 2
jinja = true
parallel = 1
swa-full = true
ctx-checkpoints = 128
stop-timeout = 600"
# Qwopus3.6-27B-Coder-MTP nothink — fine-tune coder SFT, SWE-bench 67.0%
# cache-type-v q8_0 : précision V critique pour les diffs de code
# cache-reuse 0 : incompatible MTP
MODEL_INI[qwopus3.6-27b-coder-mtp-nothink]="
model = $QWOPUS_CODER_MTP_PATH
ctx-size = 131072
cache-ram = 4096
temp = 0.7
top-k = 20
top-p = 0.8
min-p = 0.0
presence-penalty = 1.5
chat-template-kwargs = {\"enable_thinking\":false}
cache-type-v = q8_0
cache-reuse = 0
spec-type = draft-mtp
spec-draft-n-max = 2
jinja = true
parallel = 1
swa-full = true
ctx-checkpoints = 128
stop-timeout = 600"
# Gemma 4 31B — MTP drafter embarqué, texte seul (no-mmproj), spec-draft-n-max 4
# Pas de swa-full : architecture ISWA différente (issue #21468)
# chat-template-kwargs + jinja : désactive le thinking (--reasoning off non supporté)
MODEL_INI[gemma4-31b-mtp]="
model = $GEMMA_31B_PATH
spec-draft-model = $GEMMA_31B_MTP_PATH
ctx-size = 262144
cache-ram = 4096
no-mmproj = true
chat-template-kwargs = {\"enable_thinking\":false}
cache-reuse = 0
spec-type = draft-mtp
spec-draft-n-max = 4
temp = 1.0
top-k = 64
top-p = 0.95
min-p = 0.01
jinja = true
parallel = 2
stop-timeout = 600"
# Gemma 4 12B — MTP drafter embarqué, modèle unifié texte+image+audio+vidéo, texte seul
# Pas de swa-full : architecture ISWA différente (issue #21468)
MODEL_INI[gemma4-12b-mtp]="
model = $GEMMA_12B_PATH
spec-draft-model = $GEMMA_12B_MTP_PATH
ctx-size = 262144
cache-ram = 2048
no-mmproj = true
chat-template-kwargs = {\"enable_thinking\":false}
cache-reuse = 0
spec-type = draft-mtp
spec-draft-n-max = 2
temp = 1.0
top-k = 64
top-p = 0.95
min-p = 0.01
jinja = true
parallel = 3
stop-timeout = 600"
# GPT-OSS 120B — shards UD-Q4_K_XL
MODEL_INI[gpt-oss]="
model = $GPTOSS_PATH
ctx-size = 131072
cache-ram = 8192
temp = 1.0
top-k = 0
top-p = 1.0
min-p = 0.0
parallel = 1
stop-timeout = 600"
# MiniMax-M2.7 — MoE 230B Mamba hybride (10B actifs), 200K ctx, shards UD-IQ4_XS
# cache-reuse 0 : MoE incompatible — pas de swa-full (architecture Mamba, non SWA)
MODEL_INI[minimax]="
model = $MINIMAX_PATH
ctx-size = 131072
cache-ram = 8192
temp = 1.0
top-k = 40
top-p = 0.95
min-p = 0.01
cache-reuse = 0
parallel = 1
stop-timeout = 600"
# =============================================================================
# Ordre d'émission des presets dans le ini
# (declare -A ne préserve pas l'ordre d'insertion)
# =============================================================================
PRESET_ORDER=(
# always-on
qwen3.5-2b
qwen3.5-9b
qwen3.6-35b-a3b-nothink
# famille 35B A3B
qwen3.6-35b-a3b
qwen3.6-35b-a3b-mtp-nothink
qwen3-coder-next
# famille 27B
qwen3.6-27b
qwen3.6-27b-nothink
qwen3.6-27b-mtp-nothink
qwopus3.6-27b-coder-mtp-nothink
# Gemma 4
gemma4-31b-mtp
gemma4-12b-mtp
# géants
gpt-oss
minimax
)
# =============================================================================
# Génération du models.ini
# =============================================================================
generate_models_ini() {
cat <<HEADER
version = 1
; =============================================================================
; Flags globaux — appliqués à tous les presets sauf surcharge locale
; =============================================================================
[*]
n-gpu-layers = 99
cache-type-k = q8_0
cache-type-v = q4_0
flash-attn = on
prio = 2
metrics = true
slot-prompt-similarity = 0.5
cache-reuse = 4096
presence-penalty = 0.0
; =============================================================================
; always-on — load-on-startup, pas de stop-timeout
; =============================================================================
HEADER
local prev_group=""
for name in "${PRESET_ORDER[@]}"; do
# Séparateurs de groupe à partir des commentaires dans PRESET_ORDER
case "$name" in
qwen3.6-35b-a3b)
echo "; ============================================================================="; echo "; Lourds — chargés à la demande, déchargés par LRU après stop-timeout"; echo "; ============================================================================="; echo ""; echo "; --- Famille 35B A3B ---"; echo "" ;;
qwen3.6-27b)
echo "; --- Famille 27B ---"; echo "" ;;
gemma4-31b-mtp)
echo "; --- Famille Gemma 4 — pas de swa-full (architecture ISWA, issue #21468) ---"; echo "" ;;
gpt-oss)
echo "; --- Géants ---"; echo "" ;;
esac
echo "[$name]"
# Émet les lignes du preset en retirant la ligne vide initiale
echo "${MODEL_INI[$name]}" | sed '/^$/d'
echo ""
done
}
# =============================================================================
# Helpers
# =============================================================================
RED='\033[0;31m'
GREEN='\033[0;32m'
YELLOW='\033[1;33m'
NC='\033[0m'
info() { echo -e "${GREEN}[INFO]${NC} $*"; }
warn() { echo -e "${YELLOW}[WARN]${NC} $*"; }
error() { echo -e "${RED}[ERROR]${NC} $*"; exit 1; }
_dl() {
local target="$1" repo="$2"
shift 2
if [[ -f "$target" ]]; then
info "$(basename "$target") déjà présent, skip."
return
fi
info "Téléchargement $(basename "$target")..."
HF_XET_HIGH_PERFORMANCE=1 hf download "$repo" "$@" --local-dir "$(dirname "$target")"
}
_dl_shard() {
local target="$1" repo="$2" glob="$3"
local shard_dir dest_dir
shard_dir="$(dirname "$target")"
dest_dir="$(dirname "$shard_dir")"
if [[ -f "$target" ]]; then
info "$(basename "$target") déjà présent, skip."
return
fi
info "Téléchargement shards $(basename "$shard_dir")..."
HF_XET_HIGH_PERFORMANCE=1 hf download "$repo" --include "$glob" --local-dir "$dest_dir"
}
# =============================================================================
# setup
# =============================================================================
cmd_setup() {
info "Vérification des dépendances..."
PACMAN_PKGS=(curl llama.cpp-vulkan python-huggingface-hub python-hf-xet)
MISSING=()
for pkg in "${PACMAN_PKGS[@]}"; do
paru -Qi "$pkg" &>/dev/null || MISSING+=("$pkg")
done
[[ ${#MISSING[@]} -gt 0 ]] && paru -S --noconfirm "${MISSING[@]}"
command -v hf >/dev/null || error "hf introuvable"
info "Création des dossiers..."
mkdir -p "$MODELS_BASE"/{qwen3.5-2b,qwen3.5-9b,qwen3.6-35b-a3b,qwen3-coder-next,qwen3.6-27b,gemma-31b,gemma-12b,gpt-oss,qwen3.6-27b-mtp,qwen3.6-35b-a3b-mtp,minimax,qwopus3.6-27b-coder-mtp}
_dl "$QWEN35_2B_PATH" "$MODEL_QWEN35_2B_REPO" "$MODEL_QWEN35_2B_FILE"
_dl "$QWEN35_9B_PATH" "$MODEL_QWEN35_9B_REPO" "$MODEL_QWEN35_9B_FILE"
_dl "$QWEN36_35B_A3B_PATH" "$MODEL_QWEN36_35B_A3B_REPO" "$MODEL_QWEN36_35B_A3B_FILE"
_dl "$QWEN3_CODER_NEXT_PATH" "$MODEL_QWEN3_CODER_NEXT_REPO" "$MODEL_QWEN3_CODER_NEXT_FILE"
_dl "$QWEN36_27B_PATH" "$MODEL_QWEN36_27B_REPO" "$MODEL_QWEN36_27B_FILE"
_dl "$GEMMA_31B_PATH" "$MODEL_GEMMA_31B_REPO" "$MODEL_GEMMA_31B_FILE"
_dl "$GEMMA_31B_MTP_PATH" "$MODEL_GEMMA_31B_REPO" "$MODEL_GEMMA_31B_MTP_FILE"
_dl "$GEMMA_12B_PATH" "$MODEL_GEMMA_12B_REPO" "$MODEL_GEMMA_12B_FILE"
_dl "$GEMMA_12B_MTP_PATH" "$MODEL_GEMMA_12B_REPO" "$MODEL_GEMMA_12B_MTP_FILE"
_dl_shard "$GPTOSS_PATH" "$MODEL_GPTOSS_REPO" "$MODEL_GPTOSS_FILE_GLOB"
_dl "$QWEN36_27B_MTP_PATH" "$MODEL_QWEN36_27B_MTP_REPO" "$MODEL_QWEN36_27B_MTP_FILE"
_dl "$QWEN36_35B_A3B_MTP_PATH" "$MODEL_QWEN36_35B_A3B_MTP_REPO" "$MODEL_QWEN36_35B_A3B_MTP_FILE"
_dl_shard "$MINIMAX_PATH" "$MODEL_MINIMAX_REPO" "$MODEL_MINIMAX_FILE_GLOB"
_dl "$QWOPUS_CODER_MTP_PATH" "$MODEL_QWOPUS_CODER_MTP_REPO" "$MODEL_QWOPUS_CODER_MTP_FILE"
info "Génération de models.ini..."
generate_models_ini > "$CONFIG_DIR/models.ini"
info "✅ Config générée : $CONFIG_DIR/models.ini"
info "Setup terminé → ./setup-llm.sh --start"
}
# =============================================================================
# start
# =============================================================================
cmd_start() {
export PATH="$HOME/.local/bin:$PATH"
command -v llama-server >/dev/null || error "llama-server introuvable"
[[ -f "$CONFIG_DIR/models.ini" ]] || error "Config introuvable — lance d'abord --setup"
info "Lancement de llama-server (router mode) sur :8009..."
# --models-max 3 : 2 always-on + 1 lourd actif en LRU
# WebUI disponible sur http://0.0.0.0:8009
llama-server \
--host 0.0.0.0 \
--port 8009 \
--models-preset "$CONFIG_DIR/models.ini" \
--models-max 3 \
--models-autoload \
--jinja
}
# =============================================================================
# Service systemd système
# =============================================================================
SERVICE_NAME="llama-server"
SERVICE_FILE="/etc/systemd/system/${SERVICE_NAME}.service"
cmd_install_service() {
info "Génération du service systemd système..."
local script_path
script_path="$(realpath "$0")"
sudo tee "$SERVICE_FILE" > /dev/null << SERVICE
[Unit]
Description=llama-server — LLM router mode natif
After=network.target
[Service]
Type=simple
User=$USER
Group=$(id -gn)
ExecStart=${script_path} --start
Restart=on-failure
RestartSec=5
StandardOutput=journal
StandardError=journal
Environment="PATH=$HOME/.local/bin:/usr/local/bin:/usr/bin:/bin"
Environment="HOME=$HOME"
[Install]
WantedBy=multi-user.target
SERVICE
sudo systemctl daemon-reload
sudo systemctl enable "$SERVICE_NAME"
info "✅ Service installé : $SERVICE_FILE"
info " Démarrage automatique au boot activé (multi-user.target)"
info "Commandes utiles :"
info " sudo systemctl start $SERVICE_NAME"
info " sudo systemctl stop $SERVICE_NAME"
info " sudo systemctl restart $SERVICE_NAME"
info " sudo systemctl status $SERVICE_NAME"
info " journalctl -u $SERVICE_NAME -f"
}
cmd_uninstall_service() {
if ! systemctl is-enabled "$SERVICE_NAME" &>/dev/null; then
warn "Service '$SERVICE_NAME' non installé, rien à faire."
return
fi
sudo systemctl stop "$SERVICE_NAME" 2>/dev/null || true
sudo systemctl disable "$SERVICE_NAME"
sudo rm -f "$SERVICE_FILE"
sudo systemctl daemon-reload
info "✅ Service '$SERVICE_NAME' désinstallé."
}
# =============================================================================
# Entrypoint
# =============================================================================
case "${1:-}" in
--setup) cmd_setup ;;
--start | "") cmd_start ;;
--install-service) cmd_install_service ;;
--uninstall-service) cmd_uninstall_service ;;
*) error "Usage : $0 [--setup|--start|--install-service|--uninstall-service]" ;;
esac
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment