New Disk shows less Capacity

[Zonediver]

Hi folks,

i implemented a new 4TB Disk yesterday and i found some strange things now on the Main-Page of unraid:

1. The Disk has 488MB less then the other 4TB-Disk. All of them are WD-REDs

2. The Disk is empty but shows 4517MB used space

Some suggestions, why this infos are different?

 

Edited May 31, 2017 by Zonediver

[JorgeB]

Either HPA is in use or it was an external disk, you can check using the serial number on WD's RMA page.

[Zonediver]

Its a brand new disk bought yesterday.

The Disk was welded in an antistatic pack and all SMART values are Zero.

Edited May 31, 2017 by Zonediver

[JorgeB]

Tools -> Diagnostics

[Zonediver]

Sorry - agian - i forgot...

unraid-diagnostics-20170531-1117.zip

[JorgeB]

Disks are all the same size, it may be related to it being rebuild from a smaller disk, wouldn't worry about it.

 

As for the used size it's normal to have a few GBs used with an empty disk.

[Zonediver]

Just now, johnnie.black said:

Disks are all the same size, it may be related to it being rebuild from a smaller disk, wouldn't worry about it.

 

As for the used size it's normal to have a few GBs used with an empty disk.

 

Ok then i ignore it

Thanks again for your help johnnie.

[John_M]

Regarding your first point, I have noticed this in the past after replacing a disk with a larger one. The process of expanding the file system to occupy the whole of the available space doesn't appear to be as successful as if you simply add a disk and format it. I had a discussion about it couple of years or so ago and I'll add the link if I can find it. Which process did you use to introduce the new disk to the array?

 

On your second point, older versions of unRAID used to report a fairly consistent 30 or so megabytes used by an empty XFS file system, regardless of the capacity of the disk. Nowadays, it seems more like 0.11% of the disk capacity. That makes quite a significant difference with multi-terabyte hard disks. You're seeing over 4 GB used. On 5 and 6 TB disks I see 5.5 and 6.6 GB used, respectively. There was a big update to XFS around the time unRAID 6.3 appeared. I should think that was the cause. FWIW my three servers are running 6.3.4, 6.3.5 and 6.4.0-rc2 and they all show much greater disk usage by empty XFS file systems than they used to.

[JonathanM]

8 minutes ago, John_M said:

On your second point, older versions of unRAID used to report a fairly consistent 30 or so megabytes used by an empty XFS file system, regardless of the capacity of the disk. Nowadays, it seems more like 0.11% of the disk capacity

WAG, maybe the space reservation for file system growth is more aggressive now? The fastest portion of a spinner is the very first tracks, so reserving those for indexes and other housekeeping areas makes sense.

[Zonediver]

1 hour ago, John_M said:

Regarding your first point, I have noticed this in the past after replacing a disk with a larger one. The process of expanding the file system to occupy the whole of the available space doesn't appear to be as successful as if you simply add a disk and format it. I had a discussion about it couple of years or so ago and I'll add the link if I can find it. Which process did you use to introduce the new disk to the array?

 

On your second point, older versions of unRAID used to report a fairly consistent 30 or so megabytes used by an empty XFS file system, regardless of the capacity of the disk. Nowadays, it seems more like 0.11% of the disk capacity. That makes quite a significant difference with multi-terabyte hard disks. You're seeing over 4 GB used. On 5 and 6 TB disks I see 5.5 and 6.6 GB used, respectively. There was a big update to XFS around the time unRAID 6.3 appeared. I should think that was the cause. FWIW my three servers are running 6.3.4, 6.3.5 and 6.4.0-rc2 and they all show much greater disk usage by empty XFS file systems than they used to.

 

It was a replacement for a defective 3TB WD-Red.

I just swaped the Disk and this was the result.

Edited June 14, 2017 by Zonediver

[John_M]

@Zonediver That is consistent with my findings back in 2015, here

[Zonediver]

3 hours ago, John_M said:

@Zonediver That is consistent with my findings back in 2015, here

 

Hmmm... and how is this fixable?

EDIT: Found it - will try it and see what happens

EDIT2: After changing to reiserFS and back to xfs, the disk has now the full size again

Strange thing: reiserFS needs 33,6MB on the empty disk (4TB), xfs takes 4019MB of space...

 

EDIT3: Did a mistake... the disk has "not" the full size...

All 4TB disks have 3.999.810 MB - the reformated has only 3.998.833 MB 

So the procedure didn't help...

Edited June 15, 2017 by Zonediver

[John_M]

I never got round to trying Gary's suggestion as the disks were becoming flaky. I should think replacing the partition table would help recover the missing space. The easiest way to do that would be to pre-clear the disk but in order to do that it would first need to be removed from the array, which clearly has consequences as far as parity is concerned.

[John_M]

2 hours ago, Zonediver said:

Strange thing: reiserFS needs 33,6MB on the empty disk (4TB), xfs takes 4019MB of space...

 

Discussed above. XFS used to take up far less space.