[{"id":"IL-0137","title":{"en":"Training initialisation on a 60-node NVIDIA B200 cluster took more than 8 hours instead of minutes because fragmented small random I/O from all nodes saturated the metadata service of the VAST Data NFS storage","ko":"60노드 B200 학습 클러스터에서 노드들의 작은 랜덤 I/O 가 VAST 스토리지의 메타데이터 서비스를 포화시켜 학습 초기화가 8시간 넘게 걸렸다"},"org":"Lablup Inc","org_slug":"lablup-inc","date":"2026-06-15","severity":"unknown","root_cause":"capacity","url":"https://benchlet.com/incidentlib/app/cases/IL-0137","shared_components":["shared-filesystem"]},{"id":"IL-0129","title":{"en":"Throughput of the 384-GPU SmolLM3 pretraining run collapsed because shared WekaFS storage evicted dataset shards to S3, then kept dropping on local NVMe because the nanotron dataloader built an index that grew with the step count","ko":"SmolLM3 사전 학습에서 공유 Weka 스토리지가 데이터셋 샤드를 S3 로 내보내고 Nanotron 데이터로더의 인덱스가 스텝마다 커지면서, 학습을 시작한 지 몇 시간 만에 처리량이 거듭 급락했다."},"org":"Hugging Face","org_slug":"hugging-face","date":"2025-10-30","severity":"unknown","root_cause":"capacity","url":"https://benchlet.com/incidentlib/app/cases/IL-0129","shared_components":["shared-filesystem"]},{"id":"IL-0321","title":{"en":"GCP snapshots on all Patroni nodes stalled GitLab.com's primary PostgreSQL disk, Patroni failed over and 5xx spiked","ko":"GitLab.com 의 Patroni 주 데이터베이스 노드가 멈춰 장애 조치가 일어났고 쓰기 요청에서 5xx 오류가 튀었다"},"org":"GitLab","org_slug":"gitlab","date":"2021-01-16","severity":"unknown","root_cause":"capacity","url":"https://benchlet.com/incidentlib/app/cases/IL-0321","shared_components":["failover-manager"]},{"id":"IL-0320","title":{"en":"Postmark's primary MySQL disk filled but MySQL kept running, so MySQL-MMM never failed over and sending was delayed","ko":"주 MySQL 서버의 디스크가 가득 차 메일 발송, API, 웹 화면이 지연됐다"},"org":"Postmark","org_slug":"postmark","date":"2011-10-20","severity":"unknown","root_cause":"capacity","url":"https://benchlet.com/incidentlib/app/cases/IL-0320","shared_components":["failover-manager"]}]