openclaw - 💡(How to fix) Fix [Bug]: Ollama local run works, but openclaw agent --local times out (0 tokens) on Ubuntu VPS [1 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
openclaw/openclaw#62406Fetched 2026-04-08 03:04:45
View on GitHub
Comments
0
Participants
1
Timeline
1
Reactions
0
Author
Participants
Timeline (top)
labeled ×1

On Oracle Cloud Ubuntu VPS, ollama run returns quickly (after cold start), but openclaw agent --local times out for the same simple prompt.

Error Message

OpenClaw times out after minutes with stopReason=error / LLM timeout and zero token usage, no matter what timeout is set to. "stopReason": "error"

Root Cause

On Oracle Cloud Ubuntu VPS, ollama run returns quickly (after cold start), but openclaw agent --local times out for the same simple prompt.

Code Example

openclaw status --all -> no fatal issues. 


openclaw models status -> 

Config        : ~/.openclaw/openclaw.json
Agent dir     : ~/.openclaw/agents/main/agent
Default       : ollama/qwen2.5:14b
Fallbacks (0) : -
Image model   : -
Image fallbacks (0): -
Aliases (0)   : -
Configured models (1): ollama/qwen2.5:14b

Auth overview
Auth store    : ~/.openclaw/agents/main/agent/auth-profiles.json
Shell env     : off
Providers w/ OAuth/tokens (0): -
- ollama effective=profiles:~/.openclaw/agents/main/agent/auth-profiles.json | profiles=1 (oauth=0, token=0, api_key=1) | ollama:default=marker(ollama-local) | models.json=OL...EY | source=models.json: ~/.openclaw/agents/main/agent/models.json

OAuth/token status
- none




Openclaw channels status ->

Gateway reachable.
- Telegram default: enabled, configured, running, mode:polling, token:config




time ollama run qwen2.5:14b "reply with exactly: pong" -> 
pong


real    0m59.742s
user    0m0.172s
sys     0m0.081s

// Second run 

 time ollama run qwen 2.5:14b "reply with exactly: pong"
Pong!


real    0m2.040s
user    0m0.022s
sys     0m0.018s





time openclaw agent --agent main --session-id pong-test-1 --local --message "Reply with exactly: PONG" --timeout 300 --json > /tmp/pong1.log 2>&1
tail -n 80 /tmp/pong1.log   -> 
real    5m20.262s
user    0m20.966s
sys     0m2.923s
        "nonProjectContextChars": 7138
      },
      "injectedWorkspaceFiles": [
        {
          "name": "AGENTS.md",
          "path": "/home/ubuntu/.openclaw/workspace/AGENTS.md",
          "missing": false,
          "rawChars": 7809,
          "injectedChars": 465,
          "truncated": true
        },
        {
          "name": "SOUL.md",
          "path": "/home/ubuntu/.openclaw/workspace/SOUL.md",
          "missing": false,
          "rawChars": 1664,
          "injectedChars": 335,
          "truncated": true
        },
        {
          "name": "TOOLS.md",
          "path": "/home/ubuntu/.openclaw/workspace/TOOLS.md",
          "missing": false,
          "rawChars": 850,
          "injectedChars": 0,
          "truncated": true
        },
        {
          "name": "IDENTITY.md",
          "path": "/home/ubuntu/.openclaw/workspace/IDENTITY.md",
          "missing": false,
          "rawChars": 633,
          "injectedChars": 0,
          "truncated": true
        },
        {
          "name": "USER.md",
          "path": "/home/ubuntu/.openclaw/workspace/USER.md",
          "missing": false,
          "rawChars": 474,
          "injectedChars": 0,
          "truncated": true
        },
        {
          "name": "HEARTBEAT.md",
          "path": "/home/ubuntu/.openclaw/workspace/HEARTBEAT.md",
          "missing": false,
          "rawChars": 192,
          "injectedChars": 0,
          "truncated": true
        },
        {
          "name": "BOOTSTRAP.md",
          "path": "/home/ubuntu/.openclaw/workspace/BOOTSTRAP.md",
          "missing": false,
          "rawChars": 1450,
          "injectedChars": 0,
          "truncated": true
        }
      ],
      "skills": {
        "promptChars": 0,
        "entries": []
      },
      "tools": {
        "listChars": 0,
        "schemaChars": 89,
        "entries": [
          {
            "name": "session_status",
            "summaryChars": 336,
            "schemaChars": 89,
            "propertiesCount": 2
          }
        ]
      }
    },
    "stopReason": "error"
  }
}






then, a more complex question:

time ollama run qwen2.5:14b "What differentiates complicity from implicity?" -> 
It seems like there might be a mix-up in terminology as "implicity" is not
commonly used or recognized in standard English. However, I assume you're
comparing the terms "complicity" and "implicitness," which are related but
distinct concepts.

1. **Complicity**: This term refers to involvement in an activity that is
illegal or morally wrong, especially in a way that helps another person or
group succeed in their wrongdoing. Complicity implies active participation
or support for something harmful or unethical, even if it isn't direct
action and can be as subtle as silence when one ought to speak out.

2. **Implicitness**: This term relates more broadly to the idea of things
being understood or implied without being directly stated. It involves
indirect suggestion or implication that can exist in thoughts, feelings,
beliefs, intentions, motives, or in statements that do not explicitly
express these ideas but carry their meanings between lines.

To summarize:
- **Complicity** implies active involvement and support for wrongdoing.
- **Implicitness** refers to the concept of something being understood or
conveyed indirectly rather than directly stated.

If you meant a different comparison, please clarify your question!


real    2m4.702s
user    0m0.160s
sys     0m0.050s



-> and another, more tailored prompt:

time ollama run qwen2.5:14b "
You are a factual research assistant. Your task is to estimate and compare the price of pickled fish in landlocked countries.

Instructions:
- Only include countries that are geographically landlocked.
- Use realistic, evidence-based reasoning. If exact data is unavailable, infer prices using economic and supply chain factors.
- Do NOT invent precise statistics. Use ranges or relative estimates where necessary.

Method:
1. Identify a relevant set of landlocked countries across different regions and income levels.
2. Estimate the price of pickled fish in each country using factors such as:
   - Distance from coastal supply
   - Import costs and trade access
   - Local fish availability (freshwater vs imported)
   - Income level and purchasing power
   - Food preservation practices and demand
3. Rank the countries from most expensive to least expensive.

Output format (strict):
- Start with a short summary (23 sentences max).
- Then provide a ranked list (highest price → lowest price).

For each country:
- Country name
- Estimated price level (e.g., Very High / High / Moderate / Low)
- Bullet points explaining the reasoning:
  - Supply chain factors
  - Economic factors
  - Geographic constraints
  - Cultural/demand factors

Constraints:
- Be detailed but strictly factual.
- Do not include fluff or speculation without explanation.
- If uncertainty exists, explicitly state assumptions.

Begin.
"

-> 



won't paste the full output here, but time:


real    7m27.455s
user    0m0.242s
sys     0m0.235s


It doesn't appear to be an issue with the model.
RAW_BUFFERClick to expand / collapse

Bug type

Behavior bug (incorrect output/state without crash)

Beta release blocker

No

Summary

On Oracle Cloud Ubuntu VPS, ollama run returns quickly (after cold start), but openclaw agent --local times out for the same simple prompt.

Steps to reproduce

1: Have an oracle cloud hosted VPS using Ampere cpu's, 4 OCPU, 24 GB Ram. 2: Have OpenClaw 2026.4.5 (3e72c03) 3: Download any openclaw compatible model. Used qwen2.5:14b here but any should be fine. Smaller models run into the exact same behaviour. 4: Set openclaw up with valid config, and run openclaw gateway start. run openclaw gateway status --all shouldn't return any issues. 5: Run time ollama run qwen2.5:14b "reply with exactly: pong" twice, to confirm warm call is much quicker. (with whatever model you chose) 6: Run time openclaw agent --agent main --session-id pong-test-1 --local --message "Reply with exactly: PONG" --timeout 300 --json > /tmp/pong1.log 2>&1 tail -n 80 /tmp/pong1.log

Expected behavior

OpenClaw should return PONG like direct Ollama run.

Actual behavior

OpenClaw times out after minutes with stopReason=error / LLM timeout and zero token usage, no matter what timeout is set to.

OpenClaw version

OpenClaw 2026.4.5 (3e72c03)

Operating system

Ubuntu 22.04

Install method

git clone

Model

qwen2.5:14b

Provider / routing chain

openclaw -> model

Additional provider/model setup details

Logs, screenshots, and evidence

openclaw status --all -> no fatal issues. 


openclaw models status -> 

Config        : ~/.openclaw/openclaw.json
Agent dir     : ~/.openclaw/agents/main/agent
Default       : ollama/qwen2.5:14b
Fallbacks (0) : -
Image model   : -
Image fallbacks (0): -
Aliases (0)   : -
Configured models (1): ollama/qwen2.5:14b

Auth overview
Auth store    : ~/.openclaw/agents/main/agent/auth-profiles.json
Shell env     : off
Providers w/ OAuth/tokens (0): -
- ollama effective=profiles:~/.openclaw/agents/main/agent/auth-profiles.json | profiles=1 (oauth=0, token=0, api_key=1) | ollama:default=marker(ollama-local) | models.json=OL...EY | source=models.json: ~/.openclaw/agents/main/agent/models.json

OAuth/token status
- none




Openclaw channels status ->

Gateway reachable.
- Telegram default: enabled, configured, running, mode:polling, token:config




time ollama run qwen2.5:14b "reply with exactly: pong" -> 
pong


real    0m59.742s
user    0m0.172s
sys     0m0.081s

// Second run 

 time ollama run qwen 2.5:14b "reply with exactly: pong"
Pong!


real    0m2.040s
user    0m0.022s
sys     0m0.018s





time openclaw agent --agent main --session-id pong-test-1 --local --message "Reply with exactly: PONG" --timeout 300 --json > /tmp/pong1.log 2>&1
tail -n 80 /tmp/pong1.log   -> 
real    5m20.262s
user    0m20.966s
sys     0m2.923s
        "nonProjectContextChars": 7138
      },
      "injectedWorkspaceFiles": [
        {
          "name": "AGENTS.md",
          "path": "/home/ubuntu/.openclaw/workspace/AGENTS.md",
          "missing": false,
          "rawChars": 7809,
          "injectedChars": 465,
          "truncated": true
        },
        {
          "name": "SOUL.md",
          "path": "/home/ubuntu/.openclaw/workspace/SOUL.md",
          "missing": false,
          "rawChars": 1664,
          "injectedChars": 335,
          "truncated": true
        },
        {
          "name": "TOOLS.md",
          "path": "/home/ubuntu/.openclaw/workspace/TOOLS.md",
          "missing": false,
          "rawChars": 850,
          "injectedChars": 0,
          "truncated": true
        },
        {
          "name": "IDENTITY.md",
          "path": "/home/ubuntu/.openclaw/workspace/IDENTITY.md",
          "missing": false,
          "rawChars": 633,
          "injectedChars": 0,
          "truncated": true
        },
        {
          "name": "USER.md",
          "path": "/home/ubuntu/.openclaw/workspace/USER.md",
          "missing": false,
          "rawChars": 474,
          "injectedChars": 0,
          "truncated": true
        },
        {
          "name": "HEARTBEAT.md",
          "path": "/home/ubuntu/.openclaw/workspace/HEARTBEAT.md",
          "missing": false,
          "rawChars": 192,
          "injectedChars": 0,
          "truncated": true
        },
        {
          "name": "BOOTSTRAP.md",
          "path": "/home/ubuntu/.openclaw/workspace/BOOTSTRAP.md",
          "missing": false,
          "rawChars": 1450,
          "injectedChars": 0,
          "truncated": true
        }
      ],
      "skills": {
        "promptChars": 0,
        "entries": []
      },
      "tools": {
        "listChars": 0,
        "schemaChars": 89,
        "entries": [
          {
            "name": "session_status",
            "summaryChars": 336,
            "schemaChars": 89,
            "propertiesCount": 2
          }
        ]
      }
    },
    "stopReason": "error"
  }
}






then, a more complex question:

time ollama run qwen2.5:14b "What differentiates complicity from implicity?" -> 
It seems like there might be a mix-up in terminology as "implicity" is not
commonly used or recognized in standard English. However, I assume you're
comparing the terms "complicity" and "implicitness," which are related but
distinct concepts.

1. **Complicity**: This term refers to involvement in an activity that is
illegal or morally wrong, especially in a way that helps another person or
group succeed in their wrongdoing. Complicity implies active participation
or support for something harmful or unethical, even if it isn't direct
action and can be as subtle as silence when one ought to speak out.

2. **Implicitness**: This term relates more broadly to the idea of things
being understood or implied without being directly stated. It involves
indirect suggestion or implication that can exist in thoughts, feelings,
beliefs, intentions, motives, or in statements that do not explicitly
express these ideas but carry their meanings between lines.

To summarize:
- **Complicity** implies active involvement and support for wrongdoing.
- **Implicitness** refers to the concept of something being understood or
conveyed indirectly rather than directly stated.

If you meant a different comparison, please clarify your question!


real    2m4.702s
user    0m0.160s
sys     0m0.050s



-> and another, more tailored prompt:

time ollama run qwen2.5:14b "
You are a factual research assistant. Your task is to estimate and compare the price of pickled fish in landlocked countries.

Instructions:
- Only include countries that are geographically landlocked.
- Use realistic, evidence-based reasoning. If exact data is unavailable, infer prices using economic and supply chain factors.
- Do NOT invent precise statistics. Use ranges or relative estimates where necessary.

Method:
1. Identify a relevant set of landlocked countries across different regions and income levels.
2. Estimate the price of pickled fish in each country using factors such as:
   - Distance from coastal supply
   - Import costs and trade access
   - Local fish availability (freshwater vs imported)
   - Income level and purchasing power
   - Food preservation practices and demand
3. Rank the countries from most expensive to least expensive.

Output format (strict):
- Start with a short summary (2–3 sentences max).
- Then provide a ranked list (highest price → lowest price).

For each country:
- Country name
- Estimated price level (e.g., Very High / High / Moderate / Low)
- Bullet points explaining the reasoning:
  - Supply chain factors
  - Economic factors
  - Geographic constraints
  - Cultural/demand factors

Constraints:
- Be detailed but strictly factual.
- Do not include fluff or speculation without explanation.
- If uncertainty exists, explicitly state assumptions.

Begin.
"

-> 



won't paste the full output here, but time:


real    7m27.455s
user    0m0.242s
sys     0m0.235s


It doesn't appear to be an issue with the model.

Impact and severity

Affected: Not sure. Haven't seen anyone else run into this but I'm also not sure who else runs openclaw on Oracle VPS. Severity: Blocks workflows completely. Frequency: Always. Consequence: Unable to use openclaw.

Additional information

No response

extent analysis

TL;DR

The most likely fix for the OpenClaw agent timeout issue on Oracle Cloud Ubuntu VPS is to investigate and adjust the configuration or environment settings that differ between ollama run and openclaw agent commands.

Guidance

  1. Compare configurations: Verify that the configurations used by ollama run and openclaw agent are identical, especially the model settings and environment variables.
  2. Check resource utilization: Monitor the system resources (CPU, memory, disk space) during the execution of both commands to identify potential bottlenecks or limitations.
  3. Investigate timeout settings: Review the timeout settings for the openclaw agent command and consider increasing the timeout value or adjusting the underlying configuration to prevent premature timeouts.
  4. Environment and dependency check: Ensure that all dependencies and libraries required by OpenClaw are properly installed and configured on the Oracle Cloud Ubuntu VPS.

Example

No specific code snippet is provided due to the lack of explicit code references in the issue description.

Notes

The issue might be related to the specific environment or configuration of the Oracle Cloud Ubuntu VPS, which could be causing the openclaw agent command to timeout. Further investigation into the system settings, resource utilization, and configuration differences between the two commands is necessary to identify the root cause.

Recommendation

Apply a workaround by adjusting the timeout settings or environment configurations to prevent the openclaw agent command from timing out, while also investigating the underlying cause of the issue to implement a permanent fix.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

FAQ

Expected behavior

OpenClaw should return PONG like direct Ollama run.

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING