Hoping this is in the correct place. I am having a parity swap issue. I have had a data disk failure, however the spare drive I had as backup Is bigger than my parity drive (ordered by mistake!) I have a dual parity setup, not sure if that makes any difference in this situation. So I found a guide that said select the new drive in the parity slot that you would be replacing (so I just picked the oldest of the two parity drives) I then put the old parity drive in the failed data drive slot. Unraid then gave me the option to do the Parity swap and copy the data over. It all started well, I had read/write speeds of 220 on both drives, so I left them to it. I checked back half an hour later and 0 read/write speeds on both drives, but an average of 35% cpu usage, so it looked like something was going on. The progress % was at 2%. I decided to just leave it and check again this morning. Much to my frustration still at 2%, same cpu usage, same 0 read/write speeds. 

 

I have attached my diagnostic file from when I first noticed the 0 read/writes. I should also note that I have already tried things like changing Sata cables, I made sure that these two particular drives are connected directly to the MB rather than an expansion card. I have however ordered a new expansion card which has arrived this morning, so I can connect that up and try a new connection route hardware wise. I have also ordered an additional 8TB drive to match the other new drive so that the Parity can be scaled up once the swap and data rebuild is complete.

media_server-diagnostics-20231211-2259.zip

Please don't make multiple posts about the same thing, already replied to the other one, but since this was is already made lets continue here:

 

Apologies, I didn't know where was best, I have used the forum once before...

Is it just the new drive that is dropping out? I'll swap that cable again and the new Sata card just arrived so I can try connecting it to that as well just to be doubly sure of a clean new connection.

Also whilst I am here, the new HDD arrived as well. Do you know whether Unraid can do 2 parity swaps at the same time? or is it best to just installed the additional new 8TB after the swap and rebuild is done first?

47 minutes ago, Duncan9989 said:

Do you know whether Unraid can do 2 parity swaps at the same time?

It can.

What would you recommend as the best course of action, just do the swap on one, rebuild and then do the swap on the second new parity? Or just do the swap on both as I have them both to hand now and need to restart the process anyway?

5 minutes ago, Duncan9989 said:

What would you recommend as the best course of action, just do the swap on one, rebuild and then do the swap on the second new parity? Or just do the swap on both as I have them both to hand now and need to restart the process anyway?

It is risk v time balance.   Doing them at the same time is faster, but doing them individually means your array remains protected against another drive failing during the process.

Time isn't much of an issue, I would rather play it safe, especially given I don't have the best of luck with these things 🙈

Well I hate technology!!!

So I thought the Parity swap went ok, and I guess on the face of it, it did...

However when I went to move to stage 2, restart the array and rebuild the data drive, things were/are not looking good at all!

I have over 6000 errors in a matter of minutes, the read/write speed is in the KB/s rather than MB/s and this is on a brand new 8TB WD/Red Pro! 

My second new drive (N300) arrived and I have it hooked up to the array just not assigned to anything yet, what is the best course of action. Restart the whole process but parity swap on to the (n300) and return the WD drive, or is there something worth attempting on the WD red drive?

Cancel the rebuild and post the diagnostics.

Oh how I long for a simple life!

media_server-diagnostics-20231214-1206.zip

New parity disk is failing, since you have dual parity you can remove parity and rebuild using parity2 (note that if you leave it as his, it will still rebuild the disk correctly because it will use parity2 when errors are found on parity1, but it will take much longer).

Forgot to mention that you an also replace parity at the same time you do the rebuild, if you want.

Ok, so what do I actually do, before I swapped the drive places on the array menu, then when I went to start it gave me the option to swap, not rebuild or sync etc.

Do I just put the N300 in the array location of the failing WD drive, start the array and it will rebuild Parity 1 from parity 2 and Data 8 (which was parity 1) fro parity 2 as well?

1 hour ago, Duncan9989 said:

Do I just put the N300 in the array location of the failing WD drive, start the array and it will rebuild Parity 1 from parity 2 and Data 8 (which was parity 1) fro parity 2 as well?

Unassign Parity1 and disk8, start the array, stop the array, assign new parity and same disk8, start array to begin the rebuild.

when you say same disk 8, do you mean the disk 8 that is in assignment now, which was the original Parity 1 before the current swap attempt failed?

Correct

Started, lets hope this goes better than the last attempt 🤞

Now need to contact eBuyer and return the WD Red Pro. Lets see how easy they want to make it 🤣

1 hour ago, Duncan9989 said:

Started, lets hope this goes better than the last attempt 🤞

Now need to contact eBuyer and return the WD Red Pro. Lets see how easy they want to make it 🤣

good thing you did... I have never had good luck with WD red/green/black drives or any SATA drive... I have around 10 of them that failed over the 3 years I used them... Since switching to SAS drives I have had 2 failures in 7 years... and they are used ones at that(currently 20 of them are spinning)

Edited December 14, 20232 yr by mathomas3

On 12/14/2023 at 2:45 PM, JorgeB said:

Correct

Well that was a success, thank you!!!

Is there anyway to erase the old WD 8tb drive through unraid? I know it's faulty and I have no idea what if anything has copied over intact from the first attempt but I would rather wipe it before sending it back.

You can try running preclear on it.