Parity device disabled

[Deagarys]

I added some new extra drives today, my new drives were bigger than what i had so i had to make a new array and make my new drive the parity drive. I did i parity check which completed successfully

 

Yet my drive is still disabled. Why is that?
tower-diagnostics-20230730-2247.zip

Edited July 30, 2023 by Deagarys

[Veah]

Initial thought is to take array offline.  Remove the parity drive.  Add it back and be sure it has the format disk option button/warning when you enable the array.  It should begin building it.

[Deagarys]

21 minutes ago, Veah said:

Initial thought is to take array offline.  Remove the parity drive.  Add it back and be sure it has the format disk option button/warning when you enable the array.  It should begin building it.

I dont seem to get this button/warning

 

[Deagarys]

I read in another thread to stop array unassign the disk start it, stop the array re assign it and start it again.
This did not fix the issue.

I also read to stop the array, make a new array and start it.
This also did not fix my issue

 

[itimpi]

14 minutes ago, Deagarys said:

I read in another thread to stop array unassign the disk start it, stop the array re assign it and start it again.
This did not fix the issue.

This process should work and make Unraid start rebuilding the parity.    Can you please repeat the steps and at the end if it is still not fixed take new diagnostics so we can see what was happening at each stage.

[Deagarys]

7 minutes ago, itimpi said:

This process should work and make Unraid start rebuilding the parity.    Can you please repeat the steps and at the end if it is still not fixed take new diagnostics so we can see what was happening at each stage.

So this time doing it next to the start array it actually showed it will start a data rebuild.

 

But i got a error that parity disk is in error state and nothing happened. Hopefully you can see the problem in the diagnostics @itimpi

 

tower-diagnostics-20230731-0005.zip

Edited July 30, 2023 by Deagarys

[Veah]

Ok, got logged into mine and did a step by step for what I was attempting to convey.  Passing to you exactly what I intended and see. Sorry its ugly cause I'm lazy and copy/pasted from notes I was taking.

 

 ***

Stop array

Set parity drive to ‘no device’

Scroll down, see this:
Start will disable the missing disk and then bring the array on-line. Install a replacement disk as soon as possible.
Yes, I want to do this

Click Yes and then Start.

Array started with no parity.

 

Stop array again.

Assign my initial disk to the parity position.

Scroll down

Start array.  (Start will Parity-Sync and/or Data-Rebuild)

***

 

In about a day and a half, I expect my parity to be rebuilt.

Hope this helps.  It's all I got to offer.

[Deagarys]

11 minutes ago, Veah said:

Ok, got logged into mine and did a step by step for what I was attempting to convey.  Passing to you exactly what I intended and see. Sorry its ugly cause I'm lazy and copy/pasted from notes I was taking.

 

 ***

Stop array

Set parity drive to ‘no device’

Scroll down, see this:
Start will disable the missing disk and then bring the array on-line. Install a replacement disk as soon as possible.
Yes, I want to do this

Click Yes and then Start.

Array started with no parity.

 

Stop array again.

Assign my initial disk to the parity position.

Scroll down

Start array.  (Start will Parity-Sync and/or Data-Rebuild)

***

 

In about a day and a half, I expect my parity to be rebuilt.

Hope this helps.  It's all I got to offer.

Following your steps,

Once unassigned parity disk i do not get this message (i expect this is because my disk is already disabled)
I start array with unassigned parity disk, stop it again and assign the parity disk. start it again. it shows this next to the start button: (Start will Parity-Sync and/or Data-Rebuild)

I get these notifications:

 

And it does not start any rebuild process nor a parity-sync process.
 

Once again new diagnostics:

tower-diagnostics-20230731-0025.zip

 

(I really appreciate you actually doing the steps on your machine!)

Edited July 30, 2023 by Deagarys

[itimpi]

19 minutes ago, Deagarys said:

So this time doing it next to the start array it actually showed it will start a data rebuild.

 

But i got a error that parity disk is in error state and nothing happened. Hopefully you can see the problem in the diagnostics @itimpi

 

tower-diagnostics-20230731-0005.zip 145.52 kB · 3 downloads

The diagnostics show resets on the parity disk followed by write errors which is why it was disabled.

 

Are you sure your PSU is up to handling the drives?   The resets looked like they could be cabling errors (SATA or power) so. Heck that carefully.   You could run an extended SMART test on the drive to test its health.

[Veah]

I see CRC errors on Device Model:     ST4000DM004-2U9104
Serial Number:    WFN85C8K

 

That is your parity drive.  It looks like a new drive is in your future.

[Deagarys]

2 minutes ago, itimpi said:

The diagnostics show resets on the parity disk followed by write errors which is why it was disabled.

 

Are you sure your PSU is up to handling the drives?   The resets looked like they could be cabling errors (SATA or power) so. Heck that carefully.   You could run an extended SMART test on the drive to test its health.

i have a completely new 550watt psu in there i have no gpu just a i5 processor. Should be able to handle 6 disks.

[Deagarys]

Just now, Veah said:

I see CRC errors on Device Model:     ST4000DM004-2U9104
Serial Number:    WFN85C8K

 

That is your parity drive.  It looks like a new drive is in your future.

I literally bought it today

 

[Veah]

Was just looking at the drive log more... 12 power on hours and 

199 UDMA_CRC_Error_Count    -OSRCK   197   193   000    -    40;

you might have a lemon.  I dont think drives bounce back from cyclic redundancy check errors, but maybe I'm wrong. 

[Veah]

If you still have your receipt, I'd say its worth taking it back even if it is working based on that error.

[Deagarys]

A lemon? What does this mean hahaha faulty drive?

[Deagarys]

1 minute ago, Veah said:

If you still have your receipt, I'd say its worth taking it back even if it is working based on that error.

Will do that first thing tomorrow then. Is the other 4TB drive good?

[Veah]

The other looks fine.

 

Go into your diagnostic, under SMART, then open up both the ST4000DMxyz files side by side.  You will see the columns I see and you can compare them side by side.

 

*edit: CRC is basically the worst thing you can see in a HDD*

Edited July 30, 2023 by Veah

[Deagarys]

Ill deffenitly return the hdd then, 550watt psu is plenty for 6 disks right?

 

[Deagarys]

@Veah I changed the sata cable and it is in a rebuild process now. I see the CRC error so i will definitely return the disk. I can also still return the PSU if 550w is not enough, but i believe it should be enough. What is your opinion on this?

 

[Veah]

I did not check if you are running a gpu.  If not, should be ok barring a weak 12V rail that can sometimes happen when a psu gets really old.

[Deagarys]

PSU is brand new, and not running a GPU. I filled in a return form. Now its sleep time while it rebuilds Thank you for your help and ill keep you guys posted!

[trurl]

1 hour ago, Deagarys said:

I see the CRC error so i will definitely return the disk.

1 hour ago, Veah said:

I dont think drives bounce back from cyclic redundancy check errors, but maybe I'm wrong.

Yes, wrong.

 

CRCs are not drive problems, they are connection problems. No reason to think the disk is bad.

 

2 hours ago, Veah said:

Remove the parity drive.  Add it back and be sure it has the format disk option button/warning when you enable the array.

Just thought I would comment on this earlier bit.

 

Parity has no format since it has no filesystem. You can't format a disk in the parity slot.

[trurl]

1 hour ago, Veah said:

*edit: CRC is basically the worst thing you can see in a HDD*

No, just means connection problems. Many things much worse

[trurl]

Just to elaborate on UDMA CRC errors.

 

The drive firmware performs a checksum on received data, and if the checksum fails, it logs a UDMA CRC error. The data is resent so often there is no consequence, but frequent CRC errors means you need to fix the connection.

 

And, bad connections won't always result in CRC errors, since the drive receives no data to perform the checksum on.

[Veah]

If the errors clear that's great.  If they keep going up, not so great.  After the parity is done building, he can look and see where it's at.  Curious to know.