Running du -sh /
on root will show you the total disk usage from the root of the entire file-system. So that will be the disk usage for ALL devices/drives/partitions connected to the file-system root.
I've never had a RAID setup myself, so I don't know exactly how they interact with du
. But still - you have two 6Tb HDD's attached to the system.
And however they've been configured - both of these drives will be located somewhere off the root of the file-system. So despite the fact that they are in a RAID configuration, du
is probably traversing through both drives and adding the disk usage on both to the total.
So I'd imagine that's why the total usage of 7.9Tb is being reported.
It's the amount of storage used in root / plus the usage used on both of the 6Tb HDD's, plus any other storage devices that are connected.
So I'm guessing that if you half the 7.9Tb - you'll have a figure that is closer to the actual disk usage.
And from looking at your final post, which popped up as I was writing this post - In the output from df
- the 5.3Tb listing for /dev/vg1000/lv
will obviously be your RAID array.
So your 6Tb drives only actually have 5.3Tb of storage each (because it's RAID, data is mirrored on both drives, so both drives size and content should be identical). The reason it's only 5.3Tb is due to some reserved space for overheads for the file-system that is in use in that drive/partition.
e.g. used for the file-system journal, or other file-system metadata.
Of that 5.3Tb, 4Tb is used (or approximately 4Gb) so you have approximately 1.3Tb of space left on your RAID array.
So that is your actual answer.
So I think your original problem was, you were using du
in the wrong place. And despite the RAID configuration, the amount of data on both drives in the RAID array was somehow being counted twice - once per drive, yielding a usage that was much higher than you expected.
Perhaps running du -sh /dev/vg1000/lv
might yield a more correct value? But don't quote me on it. I've never set up, or used RAID before. In this case, it may even be that du
still somehow counts the space used on both of the drives in the array. In which case, you'd have to half the value reported by du
.
But from using the df
command - you can already see how much space has actually been used on the RAID array, so there's no need to use the du
command anyway!