Closed peelos closed 3 weeks ago
I managed to upgrade manually - I think the issue is the NVME location: /dev/nvme1n1 rather than nvme1
followed this guide to update manually: https://community.wd.com/t/firmware-upgrade-utility-for-linux/210120/11?u=taurus-forever
You selected "The image indicated by the Firmware Slot field is activated at the next reset." Did you reboot / reset the controller?
You selected "The image indicated by the Firmware Slot field is activated at the next reset." Did you reboot / reset the controller?
also tried option 1 and rebooted with the same effect
Intresting..probably Proxmox is doing some strange NVME virtualisation. Could you please do "wd_fw_update --info" and paste the output?
`========== Device Info ==========
Device : /dev/nvme1n1
Model : WD_BLACK SN850X 2000GB
Current fw version : 620361WD
Slot 1 readonly : False
Slot count : 2
Current slot : 2
Slots with firmware : {1: '4924458813524816438 (620331WD)', 2: '4924458826409718326 (620361WD)'}
Selected slot : -1
Activation without reset : True
Relative fw urls : []
Selected version :
Firmware url :
Activation mode : -1
Tmp fw file name :
========== Device Info ==========
Device : /dev/nvme0n1
Model : WD_BLACK SN850X 2000GB
Current fw version : 620361WD
Slot 1 readonly : False
Slot count : 2
Current slot : 1
Slots with firmware : {1: '4924458826409718326 (620361WD)'}
Selected slot : -1
Activation without reset : True
Relative fw urls : []
Selected version :
Firmware url :
Activation mode : -1
Tmp fw file name :
`
NVME drives are usually split into:
After digging into this i suspect that Proxmox does create some kind of non persistent link from /dev/nvme0n1 to the actual block device that the controller uses for firmware updates. This can lead to the illusion that the firmware update is applied when targeting /dev/nvme0n1, but it's likely being redirected to the controller behind the scenes. However, this redirection might not be persistent across reboots, explaining why the update doesn't stick.
I'll check if it is safe to always write to /dev/nvme0 instead of the namespace that is reported from nvme list.
Should work now. The nvme cli should now handle everything
Many thanks..
not sure how best to report this here.
Running proxmox on RAID 1 and the script in a virtual environmnet
appears to work but upon reboot the old firmware is still present: