Parity drive error overnight [SOLVED]

[xaq]

Hi, first and foremost apologies if something is amiss in this post, I'm very new to unraid (purchased last week)  and this is my first post in this forum. 
Unfortunately it is a failure post. Before writing any further here are the diagnostics reports.

 

[ edit: removed reports ]

Like I said, this happened overnight and I woke up to this. The 8TB drives (including the failing parity drive) are brand new, the other 3-4TB ones are sourced from my other computers. I ran a full "pre-clear" on all new and used drives before setting up the server.

 

So what I am failing to understand is: Is this drive in a "continuous error" state?
OR
"one error happened and Unraid took it offline?"

 

What is the error? is it something to warrant sending the disk back? If not how do I clear this error and start using this parity disk again?

 

Edited July 5, 2021 by xaq
solved

[trurl]

Your screenshot doesn't actually tell us the disk is offline since you clipped the part at the left that would have confirmed it. The diagnostics agree that parity has been disabled, though.

 

Unraid disables a disk when a write to it fails. Because of the failed write, the disk is out of sync and has to be rebuilt.

 

SMART for that disk looks OK. Check connections.

 

You can rebuild the disk to itself.

 

https://wiki.unraid.net/Manual/Storage_Management#Rebuilding_a_drive_onto_itself

 

 

[xaq]

Thanks for the quick reply @trurl 

 

here is the full screenshot, in the previous one I only intended to show the "error = 1" side.
Without even touching the box, I rebooted the system and unraid told me "all disks turned green"

 

So does this still mean I have to rebuild the whole parity? for a momentary write error? (from the link you've posted)

My perception (which can be entirely wrong) was: "random read/write errors happen all the time to mechanical hdd's but these are retried and corrected without user interference" 
So anytime there is a read/write error I have to rebuild the parity? 

Edited July 5, 2021 by xaq

[itimpi]

21 minutes ago, xaq said:

So anytime there is a read/write error I have to rebuild the parity

There was a write error that the normal retry mechanism was unable to correct.   In such a case you need to rebuild whatever disk reported that error.

[trurl]

1 hour ago, xaq said:

I rebooted the system and unraid told me "all disks turned green"

Do you mean there is no longer a red X next to the parity drive? Seems unlikely. The only way I can imagine that happening is if your flash drive couldn't be written when rebooting. Post new Diagnostics.

 

1 hour ago, xaq said:

anytime there is a read/write error I have to rebuild the parity?

As I said before

2 hours ago, trurl said:

Unraid disables a disk when a write to it fails. Because of the failed write, the disk is out of sync and has to be rebuilt.

It is out of sync because the failed write means parity is no longer consistent according to all the other disks.

 

So read errors will not disable a disk, though it is possible that Unraid will try to get the data for a failed read from the parity calculation with all other disks and then try to write it back to the disk. If that write fails then the disk would be disabled.

 

And, if the failed write were to a data disk instead of the parity disk, then the data disk would be the one that would be disabled and need rebuilding, because it would be out-of-sync, since parity would have been updated to allow you to recover the failed write(s) by rebuilding.

 

 

 

[xaq]

1 hour ago, trurl said:

Do you mean there is no longer a red X next to the parity drive? Seems unlikely. The only way I can imagine that happening is if your flash drive couldn't be written when rebooting. Post new Diagnostics.

 

 

No sorry, the X did not disappear, but it was no longer complaining about a write error.

I dropped the parity and re-added it. It is building the parity now 15hours 50 minutes remaining..

 

Out of curiosity do these speeds look normal to you? The first two are seagate ironwolf (‎ST8000VNZ04) and toshibas are Toshiba N300. Is there like a "community shared knowledgebase" of hdd's people own?

 

I'm asking this because I have doubts about my HBA which I bought of off amazon as "new kit including cables" and received a generic brown box delivery with no cables. I removed the heatsink (to install a fan) and noticed the thermal paste was frozen like a brick, and the heatsink pins are bent on one corner. So it is 100% used, but I don't want to return it because of the "global chip shortage" unless it is visibly faulty. 

 

Is there a way to compare speed other than plugging one hdd off the HBA and plugging it on the mobo.

[xaq]

4 hours ago, xaq said:

Out of curiosity do these speeds look normal to you? The first two are seagate ironwolf (‎ST8000VNZ04) and toshibas are Toshiba N300. Is there like a "community shared knowledgebase" of hdd's people own?

 

Answering my own question in case someone else stumbles upon this, found this excellent community App:

and its database here:

http://strangejourney.net/hddb/index.cfm?View=Drives&Vendor=Seagate&Model=ST8000DM004

 

[trurl]

5 hours ago, xaq said:

building the parity now 15hours 50 minutes remaining..

 

Out of curiosity do these speeds look normal to you?

Seems about right for 8TB, maybe a little better than I get.