Codex ate a hard drive in a resume loop
RexDouglass · x · 2026-07-21
A humorous but detailed post-mortem of Codex unexpectedly filling a disk.
The failure chain was: an orphaned codex resume --last --yolo process kept restarting from an old terminal shell, each resume forked a new rollout that recopied the full session history, and the retries created a death spiral. One loop wrote around 6 GB of history plus logs, nine copies accumulated to roughly 58 GB in four hours, and the issue went unnoticed because the disk guard was disabled and the watchdog monitored transcripts rather than filesystem growth. The author suggests a standalone disk-float watcher and a session-size hygiene rule for Codex profiles.
More from coding & agent
- Kolab Benchmark Uses Pro Screen Recordings to Show AI Agents Overestimate Collaboration Quality — minnesotanlp · 2026-10-27
- Dev argues Claude burns far more tokens than GPT models for the same coding work — haider1 · 2026-10-27
- Stop building agents that can't change models: stable contracts, external state, real evals — rseroter · 2026-10-27
- Cheap models can cost more in agent harnesses than frontier ones, researcher warns — vasuman · 2026-10-27
- Kanzo: A 5k-Line Rust Kanban Binary With MCP for Claude Code and Codex — allmudi · 2026-10-27
- 'A human approves everything' is a trap: approval fatigue and how to actually design safe agents — Efistoffeles · 2026-10-27