PUBLIC/gpu-turnstile
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases 4 Wiki Activity

Retry upstream connection-refused with exponential backoff

Browse Source 
A refused dial (service down/restarting) is retried with a wait that
doubles from BACKOFF_INITIAL (1s) up to BACKOFF_MAX (60s) until the
upstream answers or the client disconnects. Handles the Windows WSA
errno (10061) as well as POSIX ECONNREFUSED. compose.yaml.example now
uses host.docker.internal like the working local deployment.
This commit is contained in:
mram
2026-09-20 19:45:48 +02:00
parent 3aaa5d80a9
commit 20d5439b5f
6 changed files with 241 additions and 16 deletions
Download Patch File Download Diff File Expand all files Collapse all files
compose.yaml.example
+9 -3
View File
@@ -1,14 +1,20 @@
# Example deployment for gpu-turnstile. Copy to compose.yaml and adjust.
#
# gpu-turnstile listens on the ports the services normally use; the actual
# Ollama and ComfyUI instances run one port higher on the workstation.
# Ollama and ComfyUI instances run one port higher (11435 / 8189) and must
# bind 0.0.0.0 so the container can reach them (OLLAMA_HOST=0.0.0.0:11435,
# ComfyUI --listen 0.0.0.0 --port 8189).
services:
gpu-turnstile:
image: git.rambossek.at/public/gpu-turnstile:v0.1.1
restart: unless-stopped
environment:
OLLAMA_URL: http://<workstation-ip>:11435
COMFY_URL: http://<workstation-ip>:8189
# Services on the Docker host itself:
OLLAMA_URL: http://host.docker.internal:11435
COMFY_URL: http://host.docker.internal:8189
# Services on another machine: use its LAN IP instead, e.g.
# OLLAMA_URL: http://192.168.1.10:11435
# COMFY_URL: http://192.168.1.10:8189
# UNLOAD_TIMEOUT: 60s
# JOB_TIMEOUT: 15m
# LLM_WAIT_TIMEOUT: 10m