I have a share which is set to high water but the share is set to only use my 1.5TB (disk4) and 1TB (disk5) disks in my array which are the 2 smallest disks in the array (other data disks are 3 3TB).

 

Now when I add files to the share they are being written to the disk4 despite it being at 72% full where disk5 is only at 34%. Now if i understand the high-water rules correctly it takes the biggest drive capacity and divides it by 2 which "should" make 750GB then it should switch over to fill disk5.

 

Obviously this is not happening as disk4 is well over 750GB full which leads me to believe it calculates the high water based on the largest disk in the array rather than the largest disk included in the share setup, Is my assumption correct?

 

Included some pics of my setup.

Your 1TB has 654Gig free on it (which is less than the high water mark) which mean that the writes will go back to the 1.5TB until it has less than 375GB at which point the 1 TB will begin filling up to 375GB free

It will;

1 - fill disk4 until 750gig is free

2 - fill disk5 until 750gig is free

3 - fill disk4 until 375gig is free

4 - fill disk5 until 375gig is free

etc, etc.

 

Since disk5 is already under the 750gig mark it will skip step 2 and go right to step 3. Disk4 will keep getting used until it's under 375gig of free space.

 

With mismatched sizes the larger disks get used more though the first few iterations of high water. The larger the mismatch the more cycles of larger disks getting used first.

Ah of course my bad *face palm*