[SOLVED] Host crashed after removing NVMe disk

[fips]

Hey,
I have an HPE DL380 Gen10 with an NVMe active backplane and NVMe storage controller with 2x 3,84T and freshly installed Proxmox 7.4.
If I remove 1 of these NVMe disk, the server reboots after 2-3 seconds.
I did the same test with the boot disks (2x 500GB SATA SSDs ZFS Mirror) which are connected in another box through the P440 HBA Controller and the server is running happily.

Do I need to configure proxmox for NVMe hot swap?
Of course I don't want always to shut down the whole server if I have to replace a disk

Thanks

Stefan

 

[leesteken]

Adding or removing PCIe devices (like NVMe) can change the PCI ID of other PCI(e) devices. If you use PCI(e) passthrough, this can cause all kinds of problems. Also the network device name is often based on the PCI ID of the network controller, which can change and break networking.

 

[fips]

I don't have any VMs running on that fresh installed Proxmox (only installed updates).
So there is no PCI passthrough.

 

[fips]

Ups, my bad. I have overlooked the power off button on the NVMe caddys.
After switching off the drive I can safely remove it from the host