SSD Wearout not shown in GUI

[gvard]

Hello,

We use the following SSD models:

MK003840GWHTE (3.84TB HPE branded)
MTFDDAK3T8TCB (3.84TB Micron 5100 Pro)
MK001920GWHRU (1.92TB HPE branded)
MK0800GCTZB (800GB HPE branded)

The drives report SSD wearout to S.M.A.R.T., however it is on “173 (Unknown_Attribute)” (drives with 1% utilization show "99" as normalized value, with 2% utilization show "98" etc). On Micron I see 2% utilization and the value which is being used is "173 Avg_Block-Erase_Count", so 173 is being used by Proxmox in general. Is there an option to use "173 (Unknown_Attribute)" as wearout indicator too? The 3.84TB/1.92TB from HPE are Micron and the 800GB is Intel, all with HPE firmware.

 

[gvard]

On https://git.proxmox.com/?p=pve-stor...9d93038cf470fe45ba1c7168c8d0f40a5;hb=HEAD#l76 I believe I see what needs to be changed:

 402     my $vendormap = {
 403         'kingston' => 231,
 404         'samsung' => 177,
 405         'intel' => 233,
 406         'sandisk' => 233,
 407         'crucial' => 202,
 408         'default' => 233,
 409     };

I need to change 'intel' to '173' (since Intel DC S3700 use this value and are the only Intel drives I use) and change 'default' to '173'.

I know that this will get ovewritten on a future update, but:

1) Is it safe to change it?
2) Where is the file located?
3) Is there anything else which needs to be changed?

Also I'm wondering why the Micron drive shows wearout level since it is neither on the above list nor the indicator is on value 233 (it's on 173 and 202 which is for Crucial).

 

[gvard]

UPDATE: It seems the file I gave above was old, the current is https://git.proxmox.com/?p=pve-stor...12184298051d670def31b610cacce83992fa8;hb=HEAD . From there, the code which I need is:

 421     # Common register names that represent percentage values of potential
 422     # failure indicators used in drivedb.h of smartmontool's. Order matters,
 423     # as some drives may have multiple definitions
 424     my @wearoutregisters = (
 425         "Media_Wearout_Indicator",
 426         "SSD_Life_Left",
 427         "Wear_Leveling_Count",
 428         "Perc_Write\/Erase_Ct_BC",
 429         "Perc_Rated_Life_Remain",
 430         "Remaining_Lifetime_Perc",
 431         "Percent_Lifetime_Remain",
 432         "Lifetime_Left",
 433         "PCT_Life_Remaining",
 434         "Lifetime_Remaining",
 435         "Percent_Life_Remaining",
 436         "Percent_Lifetime_Used",
 437         "Perc_Rated_Life_Used"
 438     );

In SMART values, "Unknown_Attribute" is shown only 1 time and displays the value I needed, so I changed the code to:

 421     # Common register names that represent percentage values of potential
 422     # failure indicators used in drivedb.h of smartmontool's. Order matters,
 423     # as some drives may have multiple definitions
 424     my @wearoutregisters = (
 425         "Media_Wearout_Indicator",
 426         "SSD_Life_Left",
 427         "Wear_Leveling_Count",
 428         "Perc_Write\/Erase_Ct_BC",
 429         "Perc_Rated_Life_Remain",
 430         "Remaining_Lifetime_Perc",
 431         "Percent_Lifetime_Remain",
 432         "Lifetime_Left",
 433         "PCT_Life_Remaining",
 434         "Lifetime_Remaining",
 435         "Percent_Life_Remaining",
 436         "Percent_Lifetime_Used",
 437         "Unknown_Attribute",
 438         "Perc_Rated_Life_Used"
 439     );

In Micron drive, that attribute shows another value, however the script gets the value at line 431. After the change, my output is as follows:



It seemed I was lucky because that attribute is shown only once in my smartd.

Is it ok to keep that change (until it gets ovewritten in a future update)?

 

[guletz]

Hi,

First of all, I would try to upgrade the smart data base. I post a mini tutorial about this on the forum.

https://forum.proxmox.com/threads/smartdb-update-mini-tutorial.88403/

Good luck!