openclaw - 💡(How to fix) Fix [Bug]: Memory index with Ollama embeddings times out after 120s even though embeddings are ready and batch is disabled [1 comments, 2 participants]

Bug type

Crash (process/app exits or hangs)

Beta release blocker

No

Summary

Memory indexing fails with memory embeddings batch timed out after 120s when memorySearch.provider is set to ollama, even though openclaw memory status --deep reports Embeddings: ready and direct Ollama embedding requests succeed.

Steps to reproduce

1. Configure agents.defaults.memorySearch.provider = "ollama" with remote.baseUrl = "http://127.0.0.1:11434" and remote.apiKey = "ollama-local".
2. Ensure nomic-embed-text is installed in Ollama.
3. Restart the gateway.
4. Run openclaw memory status --deep and confirm it reports Provider: ollama, Model: nomic-embed-text, Embeddings: ready, and Batch: disabled.
5. Run openclaw memory index --force.
6. Observe Memory index failed (main): memory embeddings batch timed out after 120s.

Expected behavior

openclaw memory index --force should complete successfully after openclaw memory status --deep reports Provider: ollama, Embeddings: ready, and Batch: disabled, and after direct requests to Ollama's /api/embeddings endpoint succeed for the same model.

Actual behavior

openclaw memory index --force fails with Memory index failed (main): memory embeddings batch timed out after 120s. The gateway log also records memory sync failed (watch): Error: memory embeddings batch timed out after 120s.

OpenClaw version

2026.4.2 (d74a122)

Operating system

Debian Trixie

Install method

npm global

Model

openai-codex/gpt-5.4

Provider / routing chain

openclaw -> ollama (memory embeddings at http://127.0.0.1:11434)

Additional provider/model setup details

Memory search was configured with:

• agents.defaults.memorySearch.provider = "ollama"
• agents.defaults.memorySearch.remote.baseUrl = "http://127.0.0.1:11434"
• agents.defaults.memorySearch.remote.apiKey = "ollama-local"
• agents.defaults.memorySearch.remote.batch.enabled = false

Observed openclaw memory status --deep output:

• Provider: ollama (requested: ollama)
• Model: nomic-embed-text
• Embeddings: ready
• Batch: disabled

Observed ollama list output included:

• nomic-embed-text:latest

Direct embedding request that succeeded:

• curl -s http://127.0.0.1:11434/api/embeddings -d '{"model":"nomic-embed-text","prompt":"test memory embeddings"}'

Logs, screenshots, and evidence

Relevant observed log lines:
- `memory sync failed (watch): Error: memory embeddings batch timed out after 120s`
- `Memory index failed (main): memory embeddings batch timed out after 120s`

Observed `openclaw memory status --deep` output:
- `Provider: ollama (requested: ollama)`
- `Model: nomic-embed-text`
- `Embeddings: ready`
- `Batch: disabled`
- `Vector: ready`
- `FTS: ready`

Observed `ollama list` output included:
- `nomic-embed-text:latest`

Observed direct embedding API response:
- `curl` to `http://127.0.0.1:11434/api/embeddings` with model `nomic-embed-text` returned a valid embedding vector.

Impact and severity

Affected users/systems/channels: local Linux install using Ollama for memory embeddings Severity: medium Frequency: repeated across multiple observed openclaw memory index --force attempts and background memory sync attempts Consequence: semantic memory indexing does not complete, so Ollama-backed memory search cannot be used as intended

Additional information

During setup there were two earlier observed failure modes before the final reproduced state:

• openai embeddings failed: 404 404 page not found
• Ollama embeddings HTTP 404: {"error":"model \"nomic-embed-text\" not found, try pulling it first"}

After correcting those issues, the timeout persisted in the state where:

• Ollama had nomic-embed-text installed
• direct /api/embeddings requests succeeded
• openclaw memory status --deep reported Embeddings: ready
• Batch: disabled