Skip to main content
playbook · created 2026-07-09 · actor claude-cowork · confidence high · status active tags: openclaw cron dns ollama reliability

openclaw cron list shows jobs in error; openclaw cron runs --id <id> shows FailoverError: LLM request failed: DNS lookup for the provider endpoint failed (or network connection was interrupted). Same model succeeds on other jobs at other times → transient local DNS flakiness, not a provider outage. This Mac's Telegram adapters already run on sticky fallback IPs for the same reason.

Fix

  1. Per-job local-model fallback (no network/DNS needed): openclaw cron edit <id> --fallbacks "ollama/gemma4:31b"
  2. Tool-heavy jobs that hit job execution timed out (last phase: tool-execution-started): openclaw cron edit <id> --timeout-seconds 900 (default was 300).
  3. Verify: openclaw cron get <id> shows the fallback; run a no-delivery job (autoresearch-smoke) via openclaw cron run <id> and confirm status ok.

Applied 2026-07-09 to 11 jobs. Root cause (flaky resolver) still open — consider a static resolver (1.1.1.1) in macOS network settings.

Sources: ~/.openclaw/backups/cron-dns-fallback-20260709-015249/jobs-pre-edit.json