[smartmontools-support] Reading NVME logs

[Christian Franke]

Thane K. Sherrington wrote:
> Hi all,
>     Looking at NVME logs, I find them hard to understand.
>
> Take the following logs (questions inline).
>
> smartctl pre-7.4 2022-07-17 r5397 [i686-w64-mingw32-w11-21H2(64)] 
> (CircleCI)
> ...
> === START OF SMART DATA SECTION ===
> **>>>>>>>>>>>>> *SMART overall-health self-assessment test result: 
> PASSED - I assume this is good, but a perfect indicator, since I've 
> seen regular drives say the passed SMART but had read errors.*

This message only exists for consistency with ATA and SCSI output. It 
prints PASSED if and only if the "Critical Warning" byte from 
SMART/Health info is zero.

Since the early days of ATA SMART, read errors do not imply that SMART 
failure is reported.
https://www.smartmontools.org/wiki/FAQ#ATAdriveisfailingself-testsbutSMARThealthstatusisPASSED.Whatsgoingon


> ...
> SMART/Health Information (NVMe Log 0x02)

For details about this log, see for example "Figure 207" from "NVM 
Express Base Specification, revision 2.0b":
https://nvmexpress.org/developers/nvme-specification/


> Critical Warning:                   0x00

Bit 0 of this byte would be set if "the available spare capacity has 
fallen below the threshold.". Then "FAILED!" would be printed above. See 
spec for more bits.


> *>>>>>>>>>>>>> Temperature:                        79 Celsius - This 
> seems high - is that a problem?*

Possibly. If this persists, I would suggest to add more cooling.


> **>>>>>>>>>>>>> *Available Spare:                    100% - This looks 
> like 100% of the spare is free.  I assume that's a good sign?*

Yes.


> Available Spare Threshold:          10%
> **>>>>>>>>>>>>> *Percentage Used:                    16% - Does this 
> mean that 16% of the drive is used, or 16% of the spare space is used?*

"Contains a vendor specific estimate of the percentage of NVM subsystem 
life used based on the actual usage and the manufacturer’s prediction of 
NVM life. ...".
See spec for full text.


> *>>>>>>>>>>>>> **Unsafe Shutdowns:                   703 - Do unsafe 
> shutdowns matter? *

This possibly indicates that the system was not properly shut down 
frequently.


> *>>>>>>>>>>>>> **Error Information Log Entries:      3,470 -  Is this 
> a bad thing?*

Unknown. Experience shows that drive firmware of some vendors increments 
this frequently.


> Error Information (NVMe Log 0x01, 16 of 63 entries)
> **>>>>>>>>>>>>> *No Errors Logged - If there are 3470 log entries, why 
> are no errors logged?*
>

"The controller should clear this log page by removing all entries on 
power cycle and Controller Level Reset."
This differs from ATA error logs which are persistent.

Regards,
Christian