After running a parity check in which I reset the counts before running, I would expect the read values to be pretty close to each other if not exact same values for drives of the same model. Instead, the values are all over the place.

 

Of particular note is the unusually low values for drives 2 & 6. These two have historically shown low values since Dynamix was included as the GUI.

 

Counts are more along the line of expected values during a data rebuild

The difference isn't anything to worry about -- it's somehow related to the SATA/AHCI buffering setup, although I've never seen a good explanation of why they vary so much.  Nevertheless, it's common ... I have 3 UnRAID systems and all of them get wildly varying #'s of reads during parity checks.    I've asked Tom about this and it's normal ... nothing to be concerned about.

 

... note that if you do this on an older system with all IDE drives, the counts will all be the same.

 

 

Yup, not worries. Just an observation - and that someone is gonna be thinking "This looks pretty damn odd. Is it really reading my hard drives?"

Oh, forgot to mention that the counts used to match previously to beta10 I believe.

It is interesting, however, that a parity check and a parity drive rebuild gives different results when the same operation is happening on the drives - reading from one end to the other.

Agree ... despite exchanging several e-mail with Tom some time ago (year or more) I've never really understood exactly why these counts are so different.    In earlier versions (e.g. v4.7) they were identical ... and with IDE drives they're identical.    It apparently has something to do with the buffering on SATA drives (possibly AHCI related, although I haven't tried disabling AHCI to see if that changes things).

 

... but I do know it's (a) normal and (b) nothing to worry about.    I agree it's likely to cause folks to wonder why these counts are so different.