Question QNAP TR-004 DAS maximum volume size ?

[Texshooter]

AI Google says:

"The maximum volume for a single volume on a QNAP TR-004 DAS is 63 TB for non-server versions of Windows. For volumes larger than 63 TB, QNAP recommends using the Windows Server operating system"

My question is:

If I use the QNAP TR-004 storage system configured to hardware manages RAID 5 with 4 HDDs (each drive with identical TB capacity), and I use Windows 11 (not Windows Server), and I only want a DAS rig (not NAS), does the above AI Google's statement mean that the total capacity of all 4 HDDs combined cannot exceed 63TB? I want the maximum amount of storage possible for this QNAP model. And I don't want to make the mistake of buying HDDs that are too big to work with this setup.

 

[USAFRet]

These big drive sizes sound great in principle but I just don't trust them regardless are certification, warranty, price and performance. I would rather have multiple hardware devices sharing the storage capacity and depending on the data and read/write speed, data distribution across multiple devices for resilience.

Indeed.
This is a 16TB Toshiba Enterprise, from 2 yrs ago.

7 months old. Went from 0 to 14k+ bad sectors in under a week.
Tosh replaced the drive, data recovered from the backup.

 

[evermorex76]

These big drive sizes sound great in principle but I just don't trust them regardless are certification, warranty, price and performance. I would rather have multiple hardware devices sharing the storage capacity and depending on the data and read/write speed, data distribution across multiple devices for resilience.

Once you have mirrored across two drives, you're decreasing reliability as you add disks. The only reasons to use more disks are higher performance (more drives being able to read and write at once means more throughput, except with RAID 5/6) and for capacity, or as in this case lower cost because the top size drives are exceptionally expensive.

Two drives instead of one means you have double the chance of at least one drive failing, but because you're establishing the basic protection level the reliability of the data is increased. Three drives means triple, etc. The overall reliability of the array decreases when you go past two. The data itself is less protected by stretching it across many drives than it is with a mirror, because the odds of having a second drive fail are higher. (Using RAID6 or RAID10 allows for a second drive failure with no data loss, but past that you're also screwed.) Rebuilding a RAID5/6 also INCREASES the odds of a second drive failure because of the added disk activity during a rebuild, where you're hoping and praying the rebuild completes before another failure.

 

[evermorex76]

7 months old. Went from 0 to 14k+ bad sectors in under a week.
Tosh replaced the drive, data recovered from the backup.

Hard drives have a bathtub curve failure rate. They either start to fail early, well within the warranty period, or they last almost forever.

 

[USAFRet]

Rebuilding a RAID5/6 also INCREASES the odds of a second drive failure because of the added disk activity during a rebuild, where you're hoping and praying the rebuild completes before another failure.

And in my QNAP NAS, rebuild of a 4 drive RAID 5 was on the order of 2 hours per TB.

 

[USAFRet]

Hard drives have a bathtub curve failure rate. They either start to fail early, well within the warranty period, or they last almost forever.

That was just an example of an "Enterprise" drive failing early.

 

[evermorex76]

And in my QNAP NAS, rebuild of a 4 drive RAID 5 was on the order of 2 hours per TB.

Yes, RAID generally takes a long time to rebuild, and RAID5/6 are extremely slow and bandwidth-intensive as well as CPU-intensive, so the larger arrays take days to weeks. You can speed it up usually by allocating more disk time in the controller configuration (sometimes as little as 10% is used for rebuilds), but that takes performance away from the actual usage of the array. But if it's just working as a NAS with a 1Gb link, it may still be fast enough to work even with 50% of the disk time committed to rebuilds. The more drives involved, the slower the rebuild because it also has to read data from all those individual drives. Not all RAID controllers/software even reads from multiple disks at once, and RAID5 writes are generally only equal to one disk.

 

[JeffreyP55]

These big drive sizes sound great in principle but I just don't trust them regardless are certification, warranty, price and performance. I would rather have multiple hardware devices sharing the storage capacity and depending on the data and read/write speed, data distribution across multiple devices for resilience.

https://en.wikipedia.org/wiki/Perry_McCarthy