Taking disk offline not working - no valid replicas

[Marcwa19197]

Hello,

my System:
Hp Microserver Gen8
4x 4TBDisks
Freenas on latest version, having one pool as RAIDZ1.

A few days ago i got some smart errors on /dev/da0. So i ordered an replacment device.
Now i tried to take the disk offline in the pool, which brings me an errormessage "no valid replicas".
I started an Scrub, which finishe without errors a few minutes ago. I tried again taking the disk offline, but the same errormessage appeared.

What can i do now? How can i replace the faulty disk?

If i check zpool status:

I also started an new scrub, the last one i stopped.

Strange part to me: If you look at the disk which is "DEGRADED": Why is this disk not in the "raidz1" group??!

Thank for you help and best regards.

 

[Basil Hendroff]

some smart errors on /dev/da0

ada0? da0 is the boot drive on my Gen8.

 

[Marcwa19197]

Hi,
i checked this.
My disks which are members of the pool data01 are named /dev/da0, /dev/da1, /dev/da2 and /dev/da3.

/dev/ada0 is my boot pool.

Do you think my pool data01 is completely lost?
Im a bit confused, because if i look at zpool status one disk (this is the degraded one, so it muste be da0) seems to be „outside“ the raidz1.

 

[Samuel Tai]

Your pool is safe, for the moment, but if you lose another drive, it'll be unrecoverable. You'll need to backup the contents ASAP.

Once your data is safely off the system, destroy the pool. You can replace the disk then. I recommend you recreate the pool as RAIDZ2, which will allow for the loss of 2 disks before you run out of replicas. If a member disk then becomes DEGRADED, you can take if OFFLINE, pull it, and replace it.

 

[Marcwa19197]

Backup is running right now.

So you recommend destroying the pool (data loss) and recreate it?
I tried to just take the failing device offline, which brings up the "no replicas"...

 

[garm]

Freenas on latest version, having one pool as RAIDZ1.

No you do not, you have 3 drives in a raidz1 vdev and a striped single drive that is degraded. That is why you cannot offline it, you need to finish the backup and recreate the pool properly

 

[Marcwa19197]

No you do not, you have 3 drives in a raidz1 vdev and a striped single drive that is degraded. That is why you cannot offline it, you need to finish the backup and recreate the pool properly

Thank you for this information. This explained it.
So this is not a "correct" Raidz1?

Will destroy the pool after backup is finished, replace /dev/da0 and recreate raidz1 with all 4 disks.
RaidZ2 will be better, i know, but it need the space.

Thank you and best regards!

 

[garm]

Fyi: you are incredibly lucky the disk didn’t outright die, then you would loose the pool

 

[danb35]

So this is not a "correct" Raidz1?

No, it isn't. As Garm said, you created a RAIDZ1 with three disks, and have a fourth disk (likely one that you thought you could add at a later time, to turn a three-disk RAIDZ1 into a four-disk RAIDZ1) striped with that RAIDZ1. If that disk completely fails, your entire pool will fail.

 

[Marcwa19197]

Okay, i understand. Many thanks to all!!
So i guess this was a misconfiguration on my side long long time ago. (Running this machine since Freenas 9.3; 2015). But i cant remember.

So yes.. i had incredibly luck on this.
FYI: The disk which is failing was the only Seagate i have in use, all other disks (WD Red) are running fine.