[{"id":"IL-0481","title":{"en":"RoCEv2 GIDs changed after RNIC port flaps, collapsing the communication collective and crashing LLM training jobs","ko":"RoCEv2 학습 클러스터에서 RNIC 포트가 플랩한 뒤 GID 가 바뀌어 집합 통신이 깨지고 LLM 학습 작업에 크래시가 났다"},"org":"Infrawaves","org_slug":"infrawaves","date":"2026-06-01","severity":"unknown","root_cause":"software-bug","url":"https://benchlet.com/incidentlib/app/cases/IL-0481","shared_components":["network-link"]},{"id":"IL-0347","title":{"en":"About 25% of Scaleway FR-PAR-1 hypervisors lost VPC connectivity when both VPC BGP route reflectors got stuck","ko":"FR-PAR-1 가용 영역에서 하이퍼바이저 약 25% 의 VPC 연결이 11분 동안 끊겼다"},"org":"Scaleway","org_slug":"scaleway","date":"2024-10-18","severity":"unknown","root_cause":"software-bug","url":"https://benchlet.com/incidentlib/app/cases/IL-0347","shared_components":["network-switch"]},{"id":"IL-0344","title":{"en":"Linode's US-West data centers in Fremont and Santa Clara lost connectivity as the fibre circuits between them failed","ko":"Linode US-West 의 Fremont 와 Santa Clara 데이터 센터가 인터넷 연결과 두 사이트 사이 연결을 잃었다"},"org":"Linode","org_slug":"linode","date":"2023-02-11","severity":"unknown","root_cause":"software-bug","url":"https://benchlet.com/incidentlib/app/cases/IL-0344","shared_components":["network-link"]}]