SMART data, what am I looking at?

[Zepher]

I really have no idea what the numbers mean in the bottom part of the images.


This is my 128GB SSD Boot drive on my 3770K machine,

And the games/apps SSD on the same machine,

This is my Games and Apps backup drive as well as the secondary Steam location,

And one of the media drives on the same machine,

The next two drives are on my 4790K, a 512GB boot drive and a 4TB media/games/misc drive.

 

[drescherjm]

The bottom line is all drives are in good condition. Your SSDs have 86% 100% and 99% life left in them (write endurance).

 

[Zepher]

The bottom line is all drives are in good condition. Your SSDs have 86% 100% and 99% life left in them (write endurance).

Cool, so nothing to worry about. thanks.

 

[michalrz]

I concur, you have a surprisingly healthy bunch of some very hard working drives.

The lost percentage scores you see in some of them are due to a hefty amount of hours and written data.
For example, in case 1 I would worry if the following parameters were non-zero (look at the 'raw' column):
- 01, B8, BB, C3 - read errors. May or may not be relevant - different manufacturers sometimes don't follow standards.
Some drives have a lot of those but they are fine.
You can often tell it's normal by the beforementioned values being all equal. If they're all equal, and 'corrected' is among them, the drive probably did not produce a corrupt read.
- 05, C4, C5 are often related to a degraded medium, with a chance of self-repair (with or without relocating) when you perform a full surface scan, shutdown and boot again.
- AB, AC are SSD specific, could indicate bad power, bad medium or a wonky controller
- B5 looks like a great way to test partition alignment but it's not, basically irrelevant
- C7 and B7 are very often associated with connector problems. Reseat/exchange sata data cable because this can cause false errors leading to unnecessary relocations

Aside from these values, the drives collect logs with entries containing the time (offset from 0), and commands (type, address) that failed. The Linux package "smartmontools" can read them and a forum member here said the package is also available for Windows.

If you're uncertain about certain parameters and noone seems to know, you need something like HDD Guardian that will observe changes of parameters and by observing a rapid and prolonged spiking curve tendency of a parameter you can assume something's up.

A great testing strategy, often mentioned here on [H] is running 'badblocks' in destructive write mode (badblocks -wsv /dev/sdX) for like a day or more. There is also a non-destructive r/w mode.
Most Linux Live CDs have this, but beware - even the automatic boot-time discovery and automagic mounting of NTFS partitions under Linux can have unexpected consequences if there's already a MBR or other problem.