Electrical problems in Betzy GPU rack.

Minor incident High Performance Computing Betzy
2025-01-10 11:08 CET · 1 month, 1 week, 6 days, 5 hours, 23 minutes

Updates

De-escalate

Current status:
GPU nodes are still unavailable but we are working closely with vendor to get them back online. All parts have arrived onsite and we are working to schedule some maintenance work to get the parts in place.

February 4, 2025 · 08:45 CET
Update

We are awaiting technicians from Eviden, who will travel to Betzy and go through all components and check what needs to be replaced before we can power on the rack again.
Will update again when we have a time estimate on production.
Sorry for the inconvenience this is causing.

Br,
Infra Team

January 21, 2025 · 13:35 CET
Issue

There has been a small fire in the electrical system within the rack that contains Betzy accelerators(GPUS’s). This caused the whole rack to be taken out of production until further notice.
There is a scheduled downtime next week and we will update on the situation as soon as possible.

Sorry for the inconvenience this is causing.

Br,
-Infra Team

January 10, 2025 · 11:08 CET

← Back