You are not logged in.
Hello,
In the past I had no idea one could update the firmware of such things, so I had only 1 disk which ended up dying.
Now I have 2 SSDs "KINGSTON SFYRS1000G", and I set them up to use Btrfs raid1.
They were both running firmware EIFK31.6 and I noticed the new firmware EIFK31.7 was out : so I reached out to Kingston's support (their website don't provide any download link, only "SSD Manager" which runs only on Windows).
The support was actually kind enough to provide me with a link to the new firmware, as well as the small procedure to flash the new firmware:
Gain root access:
sudo -sCopy the binary zip file attached into a local directory
Extract the zip file that contains the firmware binaryExecute fw-download command to pre-stage the firmware for commit.
Note: Use the "xxxxxxxxx.bin" binary file
nvme fw-download /dev/nvme[device] --fw=[xxxxxxxxx.bin]Activate the firmware:
nvme fw-activate /dev/nvme[device] --slot=1 --action=1Reboot your system
Issue the following command to verify that the firmware has been committed correctly:
nvme id-ctrl /dev/nvme[device]
I also checked these instructions against the Arch wiki and I thought they were more or less the same, so I followed the wiki instructions with /dev/nvme0 (I left /dev/nvme1 untouched in case I ran into issues ) :
- I checked the info of my disk
# nvme id-ctrl /dev/nvme0 -H | grep Firmware
[9:9] : 0x1 Firmware Activation Notices Supported
[4:4] : 0x1 Firmware Activate Without Reset Supported
[3:1] : 0x1 Number of Firmware Slots
[0:0] : 0 Firmware Slot 1 Read/Write
- Downloaded the firmware onto the disk
nvme fw-download -f EIFK31.7.bin /dev/nvme0
- First commit (I used slot 1 because of the output of the first command I seem to understand there's only 1 slot).
nvme fw-commit -s 1 -a 0 /dev/nvme0
- Checked that the new firmware was output with :
nvme fw-log /dev/nvme0
(It was)
- Commit again (action 2, still with slot 1)
nvme fw-commit -s 1 -a 2 /dev/nvme0
- Now I tried the "reset" as stated in the wiki :
nvme reset /dev/nvme0
And it took a **really** long time during which I was unable to execute any command (such as "htop" for example), and I couldn't even Ctrl+C said htop. Then it finally gave up with an error stating something like "network dropped during reset".
Then I rebooted.
And I ended up in the emergency shell, with `dmesg` outputing Btrfs complaining about a missing disk.
I'm able to boot by adding the option `degraded` to my grub entry configuration, but /dev shows only 1 disk. `nvme list` as well...
And even my BIOS shows that my 2nd NVMe controller is empty...
I reached out to Kingston's support again, but I'm interested in understanding what's wrong with you guys.
Any help is appreciated, thank you !
Last edited by Unda (2024-10-23 08:37:31)
Offline
FYI, Kingston is advising me to try with the other disk to check if it's a coincidence of not
Of course they advise to backup the disk first, but it's still a pain, I'd need to buy another one anyway.
It appears that something may have gone wrong with the firmware update.
The only way to know for sure if this was random or not is for you to back up the content of the other SSD and then try to perform the same firmware update on the second SSD as well.
If that works, then it was just a coincidence and replacing the drive you have for a new build should fix things.
If the second SSD is rendered unusable this way also, then we need to escalate the matter to our engineers in the USA.
And then they tell me to report if they need to replace 1 or 2 disks (depending on whether the 2nd update worked or not). So that's cool, still inconvenient, but if I have no choice...
Offline