NVME drive warns on power-off; smartd reports error log entry increase

NVME drive warns on power-off; smartd reports error log entry increase

[Christian Kastner]

[-- Attachment #1: Type: text/plain, Size: 977 bytes --]

I installed a Samsung 970 EVO Plus NVME drive last year.

After having installed smartd, I now receive a notification after each
reboot, containing the following information:

> The following warning/error was logged by the smartd daemon:
> 
> Device: /dev/nvme0, number of Error Log entries increased from 195 to 196
> 
> Device info:
> Samsung SSD 970 EVO Plus 500GB, S/N:S4EVNF0M860395T, FW:2B2QEXM7, 500 GB

Furthermore, just a moment before actual power-off, a kernel warning is
printed. (This is something recent, resulting from updates to either the
kernel or systemd on my end).

This warning is not present in any logs, so I assume it is not being
persisted. I took a video of it, and attached a screenshot of the warning.

Is this the appropriate forum to report this issue? If not: could
someone point me to the appropriate forum?

Kernel version is 5.4, more specifically it is Debian's backport for the
buster release, version 5.4.19-1~bpo10+1.

Thanks,
Christian

[-- Attachment #2: foto.jpg --]
[-- Type: image/jpeg, Size: 493281 bytes --]

[-- Attachment #3: Type: text/plain, Size: 158 bytes --]

_______________________________________________
linux-nvme mailing list
linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme

Re: NVME drive warns on power-off; smartd reports error log entry increase

[Christoph Hellwig]

On Fri, Apr 17, 2020 at 02:49:52PM +0200, Christian Kastner wrote:
> I installed a Samsung 970 EVO Plus NVME drive last year.
> 
> After having installed smartd, I now receive a notification after each
> reboot, containing the following information:
> 
> > The following warning/error was logged by the smartd daemon:
> > 
> > Device: /dev/nvme0, number of Error Log entries increased from 195 to 196
> > 
> > Device info:
> > Samsung SSD 970 EVO Plus 500GB, S/N:S4EVNF0M860395T, FW:2B2QEXM7, 500 GB
> 
> Furthermore, just a moment before actual power-off, a kernel warning is
> printed. (This is something recent, resulting from updates to either the
> kernel or systemd on my end).

Can you check with the latest Linux 5.6 upstream kernel if that still
shows up?  I have a hard time figuring out what weird mix of patches
any given distro kernel has.  Alternatively report it to the debian
bug tracking system (bugs.debian.org).

For the error log:

Please run

nvme error-log /dev/nvme0

after the reboot (make sure nvme-cli is installed) which will tell us
what entry showed up in the error log.

_______________________________________________
linux-nvme mailing list
linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme

Re: NVME drive warns on power-off; smartd reports error log entry increase

[Christian Kastner]

On 2020-04-22 10:41, Christoph Hellwig wrote:
> Can you check with the latest Linux 5.6 upstream kernel if that still
> shows up?  I have a hard time figuring out what weird mix of patches
> any given distro kernel has.  Alternatively report it to the debian
> bug tracking system (bugs.debian.org).

I tried it with 5.6.6 and a olddefconfig, and I didn't see the warning
this time.

It could indeed be Debian-specific, or have been fixed after 5.4. I will
attempt to try out our own latest kernel to determine which is the case.

> For the error log:
> 
> Please run
> 
> nvme error-log /dev/nvme0
> 
> after the reboot (make sure nvme-cli is installed) which will tell us
> what entry showed up in the error log.

This never showed anything (I've tried it in the past):


Error Log Entries for device:nvme0 entries:64

.................
 Entry[ 0]
.................
error_count  : 0
sqid         : 0
cmdid        : 0
status_field : 0(SUCCESS: The command completed successfully)
parm_err_loc : 0
lba          : 0
nsid         : 0
vs           : 0
cs           : 0
.................


Oddly enough, smartctl keeps reporting an increase of what appears to be
a counter,

     Error Information Log Entries:      208

yet further down in the output, the nvme error-log output is mirrored again:

    Error Information (NVMe Log 0x01, max 64 entries)
    No Errors Logged

In any case, continuing this at bugs.debian.org as you suggested now
seems more appropriate.


Thanks,
Christian

_______________________________________________
linux-nvme mailing list
linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme