Issue: Persistent "LLM Request Timed Out" / Instant Connection Error with Local Ollama (WSL2)
Issue: Persistent "LLM Request Timed Out" / Instant Connection Error with Local Ollama (WSL2)
Describe what you're trying to do:
I am trying to run OpenClaw (v2026.3.2) on WSL2 using a local Ollama instance with the dolphin-phi model. I have successfully paired my Telegram bot (@ZenthorBot), and the gateway is running as a user systemd service.
What's happening instead:
Every message sent to the bot results in an immediate "Connection error" in Telegram and an "LLM request timed out" in the logs. Even after increasing timeoutSeconds to 300, the error remains "instant," suggesting the provider is stuck in a cooldown loop.
Relevant Config & Logs:
OS: linux 6.6.87.2-microsoft-standard-WSL2 (x64)
OpenClaw Version: 2026.3.2
Model: ollama/dolphin-phi
Gateway Status: Running (PID 25210), reachable via local loopback.
Error Logs: ```text
[telegram] [default] starting provider (@ZenthorBot)
[gateway] isError=true error="LLM request timed out"
[telegram] sendMessage ok chat=7385651919 (Sends "Connection error")
Troubleshooting steps already taken:
Increased agents.defaults.timeoutSeconds to 300.
Forced IPv4 using NODE_OPTIONS="--dns-result-order=ipv4first".
Verified Ollama is alive via curl http://127.0.0.1:11434/api/tags (returns JSON correctly).
Tried to clear cooldowns by deleting ~/.openclaw/agents/main/agent/auth-profiles.json.
Restarted via systemctl --user restart openclaw-gateway.service.
Confirmed dolphin-phi is pulled and runs fine directly in terminal via ollama run.
Additional Context:
It seems like the gateway keeps flagging the Ollama provider as "failed" and enters an immediate cooldown. Because it's on WSL2, I suspect a networking or service-communication ghost, but the auth-profiles.json wipe doesn't seem to be stopping the "instant" nature of the failure.
Describe what you're trying to do:
I am trying to run OpenClaw (v2026.3.2) on WSL2 using a local Ollama instance with the dolphin-phi model. I have successfully paired my Telegram bot (@ZenthorBot), and the gateway is running as a user systemd service.
What's happening instead:
Every message sent to the bot results in an immediate "Connection error" in Telegram and an "LLM request timed out" in the logs. Even after increasing timeoutSeconds to 300, the error remains "instant," suggesting the provider is stuck in a cooldown loop.
Relevant Config & Logs:
OS: linux 6.6.87.2-microsoft-standard-WSL2 (x64)
OpenClaw Version: 2026.3.2
Model: ollama/dolphin-phi
Gateway Status: Running (PID 25210), reachable via local loopback.
Error Logs: ```text
[telegram] [default] starting provider (@ZenthorBot)
[gateway] isError=true error="LLM request timed out"
[telegram] sendMessage ok chat=7385651919 (Sends "Connection error")
Troubleshooting steps already taken:
Increased agents.defaults.timeoutSeconds to 300.
Forced IPv4 using NODE_OPTIONS="--dns-result-order=ipv4first".
Verified Ollama is alive via curl http://127.0.0.1:11434/api/tags (returns JSON correctly).
Tried to clear cooldowns by deleting ~/.openclaw/agents/main/agent/auth-profiles.json.
Restarted via systemctl --user restart openclaw-gateway.service.
Confirmed dolphin-phi is pulled and runs fine directly in terminal via ollama run.
Additional Context:
It seems like the gateway keeps flagging the Ollama provider as "failed" and enters an immediate cooldown. Because it's on WSL2, I suspect a networking or service-communication ghost, but the auth-profiles.json wipe doesn't seem to be stopping the "instant" nature of the failure.
The AI that actually does things. Emails, calendar, home automation - all from your favorite chat app. New shell, ready to help. The lobster way.
130,529Members
Resources
Similar Threads
Was this page helpful?
