nvme format fails
eosolid opened this issue · comments
I have a drive where the format command is failing, the odd part here is that is only being observed on a dell platform. Same drive is formatting like a champ on a supermicro server. I installed the nvme-cli using the same steps on both platform and version.
this was tested on nvme 2.2.x and on nvme 2.8
CMD: nvme format --lbaf 5 --pi 0 --ms 1 --pil 0 --ses=1 /dev/nvme0n1 -t 999999999
nvme 2.2.x
nvme 2.8
Does " (0x210a) " this error code means anything to anyone??
Would it be possible that the dell platform has some sort of hardware drive lock that I should look for?
The "0x210a" is the status code the device responded with for the command. The human readable text for that status code precedes it in the output: "Invalid Format: The LBA Format specified is not supported".
As to why the device responds that way, I can't tell. With the last 3 commands succeeding, and the id-ns output snippet you're showing, it sure looks like the previous ones should have also succeeded (though I can't tell about the "ms=1" case, some devices don't support that). Sounds like a firmware or quirky device issue.
@keithbusch so as far as you are aware it... do you know if OEMs implement platform restrictions that can keep me from giving the format I wish to the drive?
The reason of the odd question is because a solution to something similar here:
https://superuser.com/questions/1574424/format-ssd-with-nvme-lba-format-specified-is-not-supported
and like I said, is very odd that is only on my dell platform that the drive fails to format, same OS and kernel on both systems.
The error status is coming from the device, so if the platform really is doing something to block this, then I assume the device cooperating with whatever mechanism the platform uses to achieve it. There are out-of-band management mechanisms for a platform to do low level device communications and configuration, so I guess it's possible even if I've never heard of any using it this way.
Yeah I need to look somewhere else... I know for a fact that is not the nvme tool ... is might as well be something in the firmware as you mention...
Thanks for your thoughts
It seems like there is a decent chance this is related to some of the other issues filed against nvme-cli for similar error messages, although I'll admit this one is a bit different (especially surprising/different is that it works just fine on a supermicro board).
Anyways, I recently figured out more details about what was going on in those related issues. In case it may help with this one, I wanted to mention the update I posted here.
Yes for me this issue was firmware related. I finally was able to complete a fw update on the device and I notice that updating the FW fix that issue.