Parity Swap - Cant Rebuild - Help Please

[lanky8804]

Am really hoping someone is able to offer some help here as stuck.

 

So my parity disk (3TB) failed and disabled itself. I have purchased a new 3TB drive to replace it with.

 

However now when i assign my new parity it will not let me rebuild and has message 'parity disk is not biggest'.

 

Have found the hpa issue but not sure if this is whats affecting mine? Have added some screenshots to hopefully help identify the issue.

 

Support greatly appreciated.

 

 

[JorgeB]

Please post your diagnostics: Tools -> Diagnostics

[lanky8804]

Diagnostics attached

unraid-diagnostics-20180620-1357.zip

[JorgeB]

Though not registered on the syslog it does look like HPA is enable on the new parity disk, it's definitely smaller than the other ones, try removing it, but because of the bad/missing sense data errors you might need to use another controller or computer to remove it, two different ways of removing it in this thread.

[lanky8804]

So just for clarity, commands to use:

 

1. hdparm -N p(?????) /dev/sd?

2. hdparm -N /dev/sde

 

What number am i adding and what sd device?

 

Thanks

[JorgeB]

Looking better at the screenshot you posted the bad/missing error is not for that disk, that disk appears to have HPA disable, but it's smaller than the other ones, which is very strange especially since they are the same model, unless that one came from an external enclosure, still no harm in trying but most likely it wont work:

 

hdparm -N p5860533168 /dev/sdb

 

Identifier might change after a reboot, check it's still sdb, you can see that on the main page, after the brand and model

[lanky8804]

so tried that and rebooted but still the same.

[JorgeB]

Yes, like I suspected the disk is actually smaller than the other ones, possibly came from an external enclosure, in any case but you'll need to get a new one, that one won't work for parity, it will for adding as a new array disk.

[lanky8804]

bummer

 

Thank you for your help.

[CaliHeatx]

On 6/20/2018 at 10:37 AM, johnnie.black said:

Yes, like I suspected the disk is actually smaller than the other ones, possibly came from an external enclosure, in any case but you'll need to get a new one, that one won't work for parity, it will for adding as a new array disk.

Yeah I've heard that some HDDs from external enclosures have a small reserved partition that makes the useable space a bit less than advertised.

[pwm]

1 hour ago, CaliHeatx said:

Yeah I've heard that some HDDs from external enclosures have a small reserved partition that makes the useable space a bit less than advertised.

 

A more important factor is that when a manufacturer claims a disk is 3 TB, they haven't actually promised the size with 12 value digits. Some disks may be 3*1000*1000*1000*1024 bytes large. Some may be 3*1000*1000*1024*1024 bytes large. Some may be 3.0001*1000*1000*1000*1024 bytes large. There is normally some fine print somewhere that describes how the manufacturer defined the size 3 TB for that specific model. Even a single manufacturer might use different definitions for different models.

[Mat1926]

29 minutes ago, pwm said:

Some may be 3.0001*1000*1000*1000*1024 bytes large. 

 

Would you please explain the reasoning for this one? What are they trying to achieve here?

 

Thnx

[JorgeB]

8 hours ago, pwm said:

A more important factor is that when a manufacturer claims a disk is 3 TB, they haven't actually promised the size with 12 value digits. Some disks may be 3*1000*1000*1000*1024 bytes large.

AFAIK, and except for external disks, for some time now, every disk from the same capacity from every manufacturer should have the same exact number of sectors, this is done to avoid issues when disks are used to replace failed RAID disks, to make sure they are always the same size.

[lanky8804]

Nothings ever simple.

 

Now gone for a 4tb just to be safe

[pwm]

3 hours ago, johnnie.black said:

AFAIK, and except for external disks, for some time now, every disk from the same capacity from every manufacturer should have the same exact number of sectors, this is done to avoid issues when disks are used to replace failed RAID disks, to make sure they are always the same size.

 

It is common that they have. Especially now when the debate if a GB should be 1024*1024*1024 or 1000*1000*1000  has been resolved and the manufacturers have all started to use the ISO definition.

 

But there isn't any law that the manufacturers needs to use the identical sector count, and the 3TB size was introduced quite a number of years ago (I bought my first 3TB disks 2010).

 

I also have disks in this size range that differs in actual number of sectors.

[CaliHeatx]

19 hours ago, pwm said:

 

A more important factor is that when a manufacturer claims a disk is 3 TB, they haven't actually promised the size with 12 value digits. Some disks may be 3*1000*1000*1000*1024 bytes large. Some may be 3*1000*1000*1024*1024 bytes large. Some may be 3.0001*1000*1000*1000*1024 bytes large. There is normally some fine print somewhere that describes how the manufacturer defined the size 3 TB for that specific model. Even a single manufacturer might use different definitions for different models.

Wow never knew this. That's ridiculous lol. Too bad there's no standard method of reporting storage capacity.

[pwm]

39 minutes ago, CaliHeatx said:

Wow never knew this. That's ridiculous lol. Too bad there's no standard method of reporting storage capacity.

 

I'm not sure if there is any formally signed standard for it now, but the manufacturers have at least unofficially locked down on the metric standard and on matching each other for "same size" drives. 10 or 15 years ago, there was a huge debate between the manufacturers and between customers and manufacturers about what the size should mean. As the differences between metric and binary-based sizes escalates the larger the disks becomes, it started to represent significant money.

 

2^10 / 10^3 = 1.024 (kB)

2^20 / 10^6 = 1.049 (MB)

2^30 / 10^9 = 1.074 (GB)

2^40 / 10^12 = 1.100 (TB)

2^50 / 10^15 = 1.126 (EB)

2^60 / 10^18 = 1.153 (PB)

 

Below figures shows that current disks have their size matching the metric size.

That the numbers aren't exact is because the sector sizes are 512 byte (or 4096 byte)

I don't have any really old drives running anymore, but below is three different 4TB disks that all has identical size reported.

WD 500 GB HDD              User Capacity:       500,107,862,016 bytes [500 GB]  = 488 386 584 * 1024
Intel 240 GB SSD           User Capacity:       240,057,409,536 bytes [240 GB]  = 234 431 064 * 1024
Samsung 256 GB SSD         User Capacity:       256,060,514,304 bytes [256 GB]  = 250 059 096 * 1024
Intel 480 GB SSD           User Capacity:       480 103 981 056 bytes [480 GB]  = 468 851 544 * 1024
WD 2.5" 1TB HDD            User Capacity:     1,000,204,886,016 bytes [1.00 TB] = 976 762 584 * 1024
Seagate Desktop 4TB HDD    User Capacity:     4 000 787 030 016 bytes [4,00 TB] = 3 907 018 584 * 1024
Seagate NAS 4TB HDD        User Capacity:     4,000,787,030,016 bytes [4.00 TB] = 3 907 018 584 * 1024
WD Red 4TB HDD             User Capacity:     4,000,787,030,016 bytes [4.00 TB] = 3 907 018 584 * 1024
WD Red 6TB HDD             User Capacity:     6 001 175 126 016 bytes [6,00 TB] = 5 860 522 584 * 1024
Seagate 10 TB HDD          User Capacity:    10,000,831,348,736 bytes [10.0 TB] = 9314 * 1024 * 1024 * 1024

 

It's a bit interesting that only the 10 TB Seagate has a size that can be evenly divided by a GiB (2^30).

All other sizes can only be evenly divided with KiB (2^10)