Hey guys,
So my existing parity drive is 6TB.
I want to replace it with a new, bigger drive (8TB), demote the ex-parity drive to an array drive, which will replace an existing 3TB drive.

Specs:
Intel® Core™ i5-7400
ASRock Z270 Gaming-ITX/ac

16GB RAM

I followed the procedure exactly as it says on the official Unraid documentation, and so far so good.

It's up to the stage where I'm copying parity info from the old drive (which has now been assigned to the array slow that the old 3TB drive used to be) to the new parity drive (assigned in the same slot as the old parity drive).

I started this process early yesterday evening at around 10.30pm.

It is now 4pm *the next day* and it's still on 26%!

Is this normal? If not, what can I do? I'm scared of stopping/interrupting anything in case I screw my array.

Here's my diags.nas-diagnostics-20221217-1603.zip

Edited December 17, 20223 yr by Corvus

Hey guys,

So I'm having a similar issue to the OP, where the parity info is taking criminally slow from my old drive to the new drive.

 

If I do a new config as recommended to the OP earlier in this thread, will I lose any data/shares on the array? 

1 hour ago, Corvus said:

will I lose any data/shares on the array?

You will lose the contents of disk1 unless you can recover it from the drive that was removed from the array.

No obvious errors but it is going pretty slow, at about 25MB/s, IIRC I've seen this issue before a couple of times because Unraid commands the disks to spin down during a parity copy, and some disks don't recover to full speed after that with just the dd operation, you could restart after disabling spin down to see if it goes any faster, or assuming the 3TB drive didn't fail do a new config with it and the new parity, sync parity then replace the drive.

4 hours ago, itimpi said:

You will lose the contents of disk1 unless you can recover it from the drive that was removed from the array.

But disk 1 was the previous parity drive. You mean I'll lose the contents of what used to be in the old disk that was in that slot?

What are my options now? At this speed, this will take a week! Why isn't this mentioned in the Unraid official documentation?? I followed it exactly.

3 hours ago, JorgeB said:

No obvious errors but it is going pretty slow, at about 25MB/s, IIRC I've seen this issue before a couple of times because Unraid commands the disks to spin down during a parity copy, and some disks don't recover to full speed after that with just the dd operation, you could restart after disabling spin down to see if it goes any faster, or assuming the 3TB drive didn't fail do a new config with it and the new parity, sync parity then replace the drive.

I doubt disabling spin down would fix it. I'm seeing all disks green right now (active), and it's still not going any faster than 25Mb/s. At this rate, it'll take a week! If I stop the procedure now, will I screw up any data/the array?

Edited December 17, 20223 yr by Corvus

Disabling spin down now won't accomplish anything, if that is the problem you'd need to disable spind down and start over.

1 hour ago, JorgeB said:

Disabling spin down now won't accomplish anything, if that is the problem you'd need to disable spind down and start over.

That's what I'm saying. The drives aren't even spun down. Every time I look at them, they're active anyway. So this will achieve nothing, *and* I'll have to start over.

So there's nothing else I can do? No idea what caused this? 

9 hours ago, Corvus said:

I'm having a similar issue

Something in an old thread that reminds you of a problem you are having is often very different in the details.

 

Also, you already have another thread going about this same problem. Please don't post about the same problem in multiple threads, it makes it impossible to coordinate responses.

 

I am going to have split your posts out of this that thread and merged them into that other this thread.

 

2 hours ago, Corvus said:

I'll lose the contents of what used to be in the old disk that was in that slot?

Do you still have that original disk1 with its contents?

18 hours ago, Corvus said:

That's what I'm saying. The drives aren't even spun down.

You misunderstood what I've posted, probably I was not very clear, Unraid will keep sending the spin down command to all disks (based on the set time) during the parity copy process, the disks will never spin down since they are active, but for some brands/models receiving the spin down command and not waking up due to a spin up or other normal user access makes some disks stay in a state that degrades their performance.

21 hours ago, trurl said:

Do you still have that original disk1 with its contents?

Yes I do. Does this open my options?

3 hours ago, JorgeB said:

You misunderstood what I've posted, probably I was not very clear, Unraid will keep sending the spin down command to all disks (based on the set time) during the parity copy process, the disks will never spin down since they are active, but for some brands/models receiving the spin down command and not waking up due to a spin up or other normal user access makes some disks stay in a state that degrades their performance.

So if I understand correctly, what you are saying is that the parity process keeps sending the spin down command, but since the disks keep receiving read commands, the disk keeps rapidly alternating between the spin down/up states, resulting in low transfer speeds?

If true, this sounds alarming to me. Wouldn't this rapid, constant switching due to conflicting information put heavy stress on the disks?

1 hour ago, Corvus said:

So if I understand correctly, what you are saying is that the parity process keeps sending the spin down command, but since the disks keep receiving read commands, the disk keeps rapidly alternating between the spin down/up states, resulting in low transfer speeds?

No that's not what I'm saying, the spin down command is sent once every time is reaches the set spin down time (Setting -> Disk Settings), but some disks don't recover full performance after that and before being spun up by a spin up command or normal user access, I don't remember if it's possible, but if the parity copy is still going and you can click you "spin up all" do it, wait 5 minutes and post new diags.

 

1 hour ago, Corvus said:

Yes I do. Does this open my options?

 

On 12/17/2022 at 9:05 AM, JorgeB said:

or assuming the 3TB drive didn't fail do a new config with it and the new parity, sync parity then replace the drive.

 

Ok so a new development. I logged onto my NAS this afternoon, and now the old parity drive has a grey square next to it. When I hover over it, it says 'new device, in standby mode (spun down)', despite the fact that it also says 'Reading' in the status, and the fact that I've already set its spin down delay to 'never'.

Despite this, it's gone up a percentage point since then, so it looks like something's still happening. The new parity drive is writing *something* at a super slow speed, but Unraid is now showing the 'read' speed of the old parity drive to be stuck at 0.0 B/s.

How is Unraid writing data to the new parity drive when it's not reading the old one - or any drive for that matter?

P.S. When I click 'spin up all', literally nothing happens. 

Edited December 19, 20223 yr by Corvus

Please post the diagnostics.

nas-diagnostics-20221219-2157.zipHere you go!

It's still going, it finished copying the old parity and is now zeroing out the new parity extra space, and it's going faster now, just wait for it to end, it should still be showing the progress.

Yeah you're right. It's now showing 91MB/s @ 80% complete. I'm guessing it will take faster to finish the last 20%?

It should be faster now.