EPYC01 is experiencing a hardware failure affecting its NVMe storage. Both NVMe drives reported read and write I/O errors, which degraded the RAID1 array and forced the virtual machine datastore into read-only mode. This has affected running virtual machines and prevented some stopped machines from starting.
Our infrastructure provider will replace the complete server barebone on the morning of 26 August 2026. This includes the hardware responsible for the PCIe and NVMe connections. They will notify us before starting the work.
After the replacement, we will verify the RAID and filesystem before starting all affected virtual machines. We expect the issue to be resolved during the morning once these checks are complete.
We apologize for the disruption.
