WD RED Smart bad info?

First of all: thank you very much Kohdee for your answer, it has made me think of some things to find some kind of explanation.

I have tried to advance a little more in this curious case.

I did what WD suggested and I ran the test with their tool. I checked the SMART, and the tool says that the SMART is OK, although I do not see the raw values ​​of each record (I attached captures WD tool vs Crystal disk info):

 /dev/sdb

/dev/sda

In what the WD wanted, I think, was to ensure that the surface of the disk is fine. I executed the quick test and the slow test and everything has given OK. However, I still have the doubt of why the hard disk hours descend at certain times. (and that crystaldiskinfo continues to confirm my theory).

I have made the decision, following the previous comment, to update the BIOS to version 1.8 and take advantage of and update to OS6 that will have a newer Smartmontool in the monitoring (I downloaded the OS4 logs).

To leave a record here, when I transferred the information when installing OS6 it was approximately 6 days the first hard disk running without the other, to fill the volume and force a resynchronization.

That is to say, on January 25, 2017 at 7:52 pm, when I inserted the second hard disk, the statistics of the disk are:

/ dev / sda: 2929 hours
/ dev / sdb: 2643 hours

From that moment on they will be 24 hours continuously, so that the SMART hour counters of both disks should increase correctly in the future (if these counters are lower than these values ​​in the future it is definitely a SMART error that the WD tool is not able to notify).

I know that OS6 is not officially supported, but this happened on OS4 (which is official) being connected to the SATA ports of the NAS chipset (I think it's an Intel® NM10 Express) and that it should not be incompatible and now with a OS / Debian / Linux Kernel / Smartmontools much more updated.

I will keep you all informed.

Thanks to everyone.

/dev/sda/dev/sdb