on a linux box you should be able to run:
sudo fdisk /dev/sdwhatever
g to make a new blank GPT partition table
w to write this to disk
There is nothing like that. I’ve attached the pictures above of diskpart, Disk Management and there is no hidden partition or any tables. I thought the same and did the following
diskpart
list disk
select disk 4
clean
Went to Disk Management and initialized the disk in GPT and created a NTFS partition but the usable capacity reported is the same everywhere. The same can be seen in smartctl as well.
@RetroG Were you able to check with your drive?
So, @Johnny_Fartpants Here’s what i did:
- Took the brand new sealed drive (Seagate EXOS X16 16TB SN03 firmware) and hook it to the system via onboard PCH (Intel) SATA Controller
- Used Seagate Download Center and used the Serial of this new drive to obtain the firmware and the utility.
- Created a SeaChest Bootable from the zip i obtained and also copied the new firmware to the root of the USB.
- Booted using the USB on the target computer with this brand new drive
- Scanned and checked the drive firmware using SeaChest. Ran smartcl and it reported correct 16TB capacity and firmware to be SN03 and the format as 512B.
- Went ahead and flashed the firmware and the update was successful.
- Scanned using SeaChest and the firmware was showing updated to SN04. I also verified the firmware, capacity and sector size via smartctl and they were all good
- Shut down the target computer
- Waited for like 5-10mins
- Started the computer again and booted off the SeaChest USB
- Scanned using SeaChest. Did a grep size of sectors. Checked supported formats of the target disk. I also ran smartctl to verify the firmware, capacity and sector size after power cycle and they were all good
- Did a short DST test on the target disk
- Changed the sector size to 4096. This took a few minutes
- Scanned using SeaChest. I also verified the firmware, capacity and sector size via smartctl and they were all good. Did a grep size of sectors as well
- Shut Down the computer
- Waited for like 5-10mins
- Started the computer again and booted off the SeaChest USB
- Scanned using SeaChest and did a short DST test on the target disk.
- Did a grep size of sectors. Checked supported formats of the target disk. I also ran smartctl to verify the firmware, capacity and sector size after power cycle and guess what? This time, no loss in usable capacity for this particular disk!
- I also verified it using HD Sentinel, HD Tune Pro, SeaTools, diskpart and Disk Management and it shows all good. 14902.0GB as a usable space for 16TB disk!
Please note that everything was done in the same order (DITTO) for the other drive, except for the factor that the other (previous) disk had one single NTFS partition with the GPT intialization, created using Disk Management from Windows. Also, @RetroG the other day you told about the reserved space, could we conclude the following:
- My other drive had some sort of manufacturing issue or some other things or it just did not went well
- The loss of usable capacity (500GB approx) was due to the disk not being unallocated and an active partition created and when the sector sized got changed, it was unable to wipe all partition on the disk and didn’t work well.
If its the second case, then it means one has to clean/wipe all the partitions and then change the sector size or there will be a loss of usable space when changing the sector size to 4096 or vice-versa (back to 512B)
Let me know what you and @winnielinnie guys think.
Also, one thing i’m not sure of is Head_Flying_Hours. What is this thing?
Another weird thing i noticed after changing sector size to 4096 is the delay in loading Disk Information in SeaTools Interface on the previous disk and this new disk i updated the firmware+changed sector size today.
Interestingly, the SeaTools was not slow with the same disks before and the only different thing was these two disks were 512e. When changed to 4K, both the disks takes like 2-3mintues to complete scanning and then show up in SeaTools otherwise SeaTools just keeps loading for 2-3 minutes. Could this also result in slow drives in ZFS or when doing read/write to or from the pool?
not yet… I realized that I really need to redo my snapraid parity on that system (due to the potential of the drive size changing there is a risk in losing the data… I’d need to have valid parity which it is not (yet) done syncing)
another thing that makes me wonder, what happens if you convert the problem disk back to 512e
1 Like
Ouch. I got success with another drive. Have posted the details above. Please have a read and let me know your thoughts.
Oh yeah, i’ll be doing that today and report the results!
@Johnny_Fartpants Any insights?
Have you tried wiping this drive within TrueNAS to see if that changes anything? Or even build a pool with it and blow it away?
Nah, not tried with TrueNAS but have tried every wiping tool until now. What do you think what could have may caused it?
Also, one thing i’m not sure of is Head_Flying_Hours. What is this thing?
Another weird thing i noticed after changing sector size to 4096 is the delay in loading Disk Information in SeaTools Interface on the previous disk and this new disk i updated the firmware+changed sector size today.
Interestingly, the SeaTools was not slow with the same disks before and the only different thing was these two disks were 512e. When changed to 4K, both the disks takes like 2-3mintues to complete scanning and then show up in SeaTools otherwise SeaTools just keeps loading for 2-3 minutes. Could this also result in slow drives in ZFS or when doing read/write to or from the pool?
As asked by @RetroG i’ll be doing back to 512e and then check how is the result.
You might want to contact Seagate at this point. It’s odd that you would supposedly have “less” total capacity when there is less overhead using the native sector size. Then for Seagate’s own tool to act very slow just because a drive is using its native sector size?
I did not experience any of this with my Exos X20s.
I simply upgraded the firmware and did the lowlevel format to 4Kn for my drives. Then booted into TrueNAS to create a vdev/pool.
It’s the time the drive heads have been active to my knowledge.
Hmm. So no one has idea why the usable capacity was less than expected on one of the drive? I did the ditto steps for another drive and it went all good. So, what we can conclude from this incident?
- My other drive had some sort of manufacturing issue or some other things or it just did not went well
- The loss of usable capacity (500GB approx) was due to the disk not being unallocated and an active partition created and when the sector sized got changed, it was unable to wipe all partition on the disk and didn’t work well.
Yes, 100% right. Both of the drives, previous one (less capacity) and one with right capacity, both went slow in SeaTools. The SeaTools interface takes like upto 3 mins to load the interface. Whereas, both of these drives, when were in 512e, the SeaTools interface loaded so much quick, in about 30-40 seconds max.
I see so much numbers on for this particular attribute!
Where did you purchase these drives?
It was from one of a local store and i bought from two different stores to have different batches to have less failures.
Please note that the reduction in the usable capacity for one of the drive was only after updating the firmware+changing the sector size to 4K.
Brand new? Refurbished? Used?
What’s the number?
Brand new, fully sealed drives. No refurbished, not used.
For one of the disk, after creating a partition, testing it with some files, updating the firmware to SN04 and changing the sector size to 4K:
And for one of the drive, its 886199500039
I can get you the exact number for the second disk tonight!
That shows 0 hours.
What about the Power_On_Hours?
EDIT:
Stop doing that! 
That’s the number I was asking about. Not sure why you showed me the drive with a value of 0 first, and then later edit the post to say “BTW here’s a really big number of a different drive”.
Are you sure that’s the number? That’s too many hours for a drive to exist. It’s a nonsense number.