Strav's blog

Normally a load of SATA CRC error counts would be concerning, however in some cases it's not.

UPDATE: For future reference I have spoken with Kingston and it turns out that this isn't actually a CRC error rate issue. The cables are all fine. Basically, there is no standardisation for SMART data for SATA SSDs, only with M.2 NVME SSDs. Here is an email from Kingston confirming such:

Hello,

Basically it comes down to the capability of the SSD, starting from the type of controller and NAND flash used, but also including the level and type of cache available. 

The SSD controllers used in SA400S37/480G drives are designed for the SSDs to be used in standard desktops and laptops only. However, in this case, although you are using the hardware like a server, it is not a dedicated server system, and so the above does not really apply. The ASUS PRIME Z270-A is a standard desktop board from 2017 which can accommodate SA400S37/480G drives: https://www.kingston.com/memory/search/model/95749/asus-prime-z270-a-motherboard 

That being said, your initial query was regarding the value interpreted by the UNRAID software as standing for "SATA CRC Error". In fact, this is not the case and the reason for this false positive is because SMART diagnostic applications for SATA SSDs are not standardised as opposed to M.2 NVMe SSDs. This means that certain values can be incorrectly interpreted by different software applications depending on where these values are located.

For the SA400S37, value 199 is in fact vendor specific and by that we mean it can and does vary from one SA400S37/480G SSD controller to another. One uses the value to report raw data, the other SSD does not utilise the value at all. 

The SA400S37/480G uses well over 200 different builds, most of which have their own firmware. This is why you have one SSD with firmware SAI20103 and another with firmware SHFK60.1. Because there are so many builds, some SMART values will vary between one revision and another.

Going forward, this means that the only reliable way to determine the health status of the SSD(s) is to use our own dedicated Kingston SSD manager application. Using any other software is likely to result in some values being interpreted incorrectly.

Kind regards,