Layer-2 loop contained in STO, row B

时间线

1. 已解决2026年2月12日 10:03 UTC

   Row B has been clean since 10:03 UTC. Duration: 9 minutes. Affected machines receive the SLA extension automatically.

2. 已确定原因2026年2月12日 09:59 UTC

   A patching error during a machine replacement created a loop between two top-of-rack switches in row B. Storm control tripped on one side but not the other. The offending cable has been pulled.

3. 调查中2026年2月12日 09:55 UTC

   Machines in row B of STO are seeing severe packet loss since 09:54 UTC. The pattern looks like a broadcast storm. We are isolating the row.

事故复盘会在解决后 5 天内发布于此。