{"rewrite":{"id":"r_a2d939d44f42529f724104eb","clusterId":"c_f3a932f0e555c66b261707ea","slug":"codex-logging-bug-caused-up-to-640-tb-year-of-ssd-writes-mostly-fixed","model":"deepseek-v4-flash","headline":"Codex Logging Bug Caused Up To 640 TB/Year Of SSD Writes, Mostly Fixed","summary":"OpenAI's Codex AI coding assistant had a logging bug that could write approximately 640 TB of data per year to a local SSD, potentially exceeding the endurance of a typical 1 TB drive in under a year. The issue was reported by GitHub user 1996fanrui, who observed about 37 TB written over 21 days of use, with Codex SQLite logs as the primary source. Investigation showed the database retained only about 680,000 logs, but the SQLite row ID exceeded 5.5 billion, indicating repeated cycles of inserting large amounts of logs and immediately deleting them. Most writes came from TRACE level detailed logs recording function calls and variable values. OpenAI released version 81150 of win-codex a few days after the report. Thibault Soti, OpenAI's head of core products, said on X: \"There was a bug in Codex. Fixed it.\" GitHub user wfy-op reported that before the update there were 28,000 to 46,000 writes in 30 seconds, dropping to about 6,700 after the update. OpenAI claims the latest version reduces write volume by about 85%. However, TRACE level logs are still saved to the database, and wfy-op noted write volume, while reduced, is still not normal low-frequency log behavior.","whyItMatters":"The bug exposed a real-world risk of rapid SSD wear from AI agent logging, and the fix shows OpenAI can respond quickly to infrastructure-level issues reported by the user community.","webCardHtml":"\u003cp\u003eThe bug was reported in a GitHub issue on the openai/codex repository. The reporter, 1996fanrui, measured approximately 36,211 new log rows written in a 15-second window, with rows being inserted, indexed, written to the WAL, and then deleted continuously while the retention count stayed unchanged. The analysis traced the majority of writes to TRACE level logs that record function calls and variable values, along with a large volume of internal file monitoring logs.\u003c/p\u003e\n\u003cp\u003eThe report suggested several changes: not persisting TRACE logs by default, setting database size limits, and saving only the minimum necessary data instead of full WebSocket payloads. OpenAI released version 81150 of win-codex a few days later. Thibault Soti, OpenAI\u0026#39;s head of core products overseeing both ChatGPT and Codex, said on X: \u0026#34;There was a bug in Codex. Fixed it.\u0026#34;\u003c/p\u003e\n\u003cp\u003eGitHub user wfy-op reported that before the update there were 28,000 to 46,000 writes in 30 seconds, dropping to about 6,700 after the update. OpenAI claims the latest version reduces write volume by about 85%. However, TRACE level logs are still saved to the database, and wfy-op noted write volume, while reduced, is still not normal low-frequency log behavior.\u003c/p\u003e","blueskyPost":"Codex's logging bug wrote 37 TB in 21 days, enough to wear out a typical 1 TB SSD in under a year. The 85% fix still leaves TRACE-level logging active, meaning the drive wear is reduced but not eliminated.","twitterPost":"Codex's logging bug could kill a 1 TB SSD in under a year. The 85% fix still keeps TRACE logs, so drive wear continues.","threadsPost":"Codex's logging bug wrote 37 TB to a local SSD in 21 days, enough to exceed a typical 1 TB drive's endurance in under a year. OpenAI's fix reduces write volume by 85%, but TRACE level logs are still saved. The drive wear is lower but not gone.","newsletterBlurb":"A bug in OpenAI's Codex caused excessive SQLite logging that could write about 640 TB per year to a local SSD, potentially wearing out a typical 1 TB drive in under a year. OpenAI released a fix in version 81150 that reduces write volume by roughly 85%, though some logging remains.","attributionJson":"[{\"source\":\"GIGAZINE\",\"url\":\"https://gigazine.net/news/20260625-codex-ssd-write/\",\"title\":\"Codex Logging Bug Causes ~640 TB/Year of Writes to Local SSD, Mostly Fixed in Latest Version\"}]","lintFlagsJson":null,"lintHits":0,"costUsd":0,"inputTokens":11215,"outputTokens":1166,"status":"published","repairAttempts":0,"nextRepairAt":null,"factsAttemptedAt":1782522750,"createdAt":"2026-06-27T01:09:18.000Z","publishedAt":"2026-06-27T01:10:24.000Z","updatedAt":"2026-06-27T01:10:24.000Z"},"cluster":{"id":"c_f3a932f0e555c66b261707ea","canonicalTitle":"Codexのログ記録バグによりローカルSSDに年間640TB相当の書き込みを行う不具合、最新版で大半を修正","representativeArticleId":"a_184dc608a946649478f57be0","sourceCount":1,"writtenSourceCount":1,"writeAttempts":0,"isSolo":false,"entitiesJson":"{\"anime_titles\":[],\"manga_titles\":[],\"work_titles\":[],\"studios\":[],\"people\":[],\"type\":\"news\",\"domain\":\"other\",\"is_roundup\":false}","contentType":"news","status":"published","firstSeenAt":"2026-06-25T10:00:00.000Z","lastSeenAt":"2026-06-26T02:22:00.000Z","updatedAt":"2026-06-27T01:10:25.000Z"},"attribution":[{"source":"GIGAZINE","url":"https://gigazine.net/news/20260626-openai-employees-codex/","title":"OpenAI社員がAIエージェントをどれだけ使っているのかが丸わかりなレポートが公開される、非エンジニアもCodexをバリバリ活用"}],"entities":{"anime_titles":[],"manga_titles":[],"work_titles":[],"studios":[],"people":[],"type":"news","domain":"other","is_roundup":false},"keyFacts":["A bug in OpenAI's Codex AI coding assistant could write approximately 640 TB of data per year to a local SSD, potentially exceeding the endurance of a typical 1 TB drive in under a year.","GitHub user 1996fanrui reported the bug and observed about 37 TB written over 21 days of use, with Codex SQLite logs as the primary source.","OpenAI released version 81150 of win-codex a few days after the report, and the company claims the update reduces write volume by about 85%.","GitHub user wfy-op reported that before the update there were 28,000 to 46,000 writes in 30 seconds, dropping to about 6,700 after the update.","Thibault Soti, OpenAI's head of core products, said on X: 'There was a bug in Codex. Fixed it.'"]}