Detached zfs volume, no snapshot newbie

[pldowden]

I'm an idiot, I have a raid 10 setup, Freenas 9.2.0 with 6x 3TB and 2x 4TB drives, one of the 4TB failed and I tried to replace it but had no luck. The long and the short of it, is in misreading a forum, I detached the volume without any snapshots in place and now, when I try to re-import the volume, it recognizes it (auto import has it as an option with original name and ID number) but fails.

Using the Shell, if I try importing it gives the "one or more devices is currently unavailable"

I've tried cloning the bad drive and repairing it, and although the bad/repaired HDD is an option if I want to build a new volume, I still get the "one or more devices is currently unavailable" error on import attempts.

I've tried -D and -F and as a newbie, am now at a loss as to what to do.

Thought I'd ask here before I just write off the data and start from scratch.

Thanks in advance for your patience with my ignorance.

Paul

 

[Heracles]

I have a raid 10 setup

Warning here.... Give us the complete hardware details of that setup please.

Freenas 9.2.0

Second warning here... Pretty old version that should have been updated long time ago...

I detached the volume without any snapshots in place

Snapshots are themselves inside the pool, so by detaching the pool, you are detaching the snapshots anyway...

So stop playing trial N error and tell us everything about the hardware first.

 

[pldowden]

Warning here.... Give us the complete hardware details of that setup please.



Second warning here... Pretty old version that should have been updated long time ago...



Snapshots are themselves inside the pool, so by detaching the pool, you are detaching the snapshots anyway...

So stop playing trial N error and tell us everything about the hardware first.

Apologies, wasn't aware the hardware config played a part.
Running this on an HP Proliant Microserver G7, 4x 3TB SATA in the bays, 2x 3TB external USB drives, 2x 4TB external USB drives. ...does this help?

 

[Heracles]

does this help?

It does but the most important parts are missing...

HP Proliant Microserver G7

RAM ?

4x 3TB SATA in the bays,

How are these drives connected to / managed by the server ? On a hardware Raid controller ? An HBA ? ...

2x 3TB external USB drives, 2x 4TB external USB drives

Big No-No here. Never use USB drives with TrueNAS. TrueNAS must have direct and complete physical control of its all its drive.

So should we ever be able to bring that pool back online, first thing to do will be an emergency backup because that pool will have to be destroyed and re-built USB-free.

So about backups, do you already have one ? Is it good enough to forget about the actual pool and re-start from scratch ?

 

[Patrick M. Hausen]

No ... precise zpool topology, please. Actions taken and commands entered. Literally and complete. There is no RAID10 in ZFS, for example. What exactly was that what you were running?

What do you mean by "detached the volume"? There are no volumes in ZFS. If I remember correctly, FreeNAS 9.x falsely named pools "volumes". OK. But you cannot detach a pool. You can detach a disk/partition, though. So we are a bit at a loss what you did. Really.

And as I use to say, my crystal ball is at the shop for maintenance. Not intending to be snarky - simply not possible to help, because from your initial description I don't get a clue what the heck is going on.

 

[pldowden]

It does but the most important parts are missing...



RAM ?



How are these drives connected to / managed by the server ? On a hardware Raid controller ? An HBA ? ...



Big No-No here. Never use USB drives with TrueNAS. TrueNAS must have direct and complete physical control of its all its drive.

So should we ever be able to bring that pool back online, first thing to do will be an emergency backup because that pool will have to be destroyed and re-built USB-free.

So about backups, do you already have one ? Is it good enough to forget about the actual pool and re-start from scratch ?

8gb 1333
Embedded AMD SATA controller with RAID 0, 1

If I toast it, I am only losing 4gb of non-backed up data I think, albeit I don't know what those 4gb consist of, so was looking for simple solutions first.

This 1+0 raid was initially for data protection but now I have (almost) everything backed up to a cloud, I'm more interested in speed and not losing the additional 14tb of External USB storage space (mounted and accessible through the Freenas as a convenient single share).

Your learned advice is appreciated!

 

[Heracles]

So you got many things very wrong here...

USB is a major No-Go, just as hardware raid is. We may try to recover but because you said that you have a good backup, up to you to decide if it worth the trouble.

Usually the drives used in external casing are not meant for NAS usage. NAS drives are not named as such just to justify a higher price. They are different.

Should a desktop have problems reading a sector, it will fight it forever to recover it because it knows that either it get it back or it is lost.

A NAS drive will try but if it fails, it will quickly returns an error instead because that context expects redundancy. TrueNAS will then compensate for the data from the other drives and then will go back trying to fix that sector.

A desktop drive is meant to be used at most a few hours at a time, alone.
NAS drives are meant to run 24 / 7, sustain the heat and the vibration not only from themselves but from other drives next to them.

A desktop drive will try to park itself often to save energy.
NAS drive are meant to be used all the time (and they are), so will avoid spining up and down all the time to avoid accelerating their hardware degradation.

8 G of RAM is the very minimum for TrueNAS. Below that, it is known that weird problems appears and the system becomes unstable.

So about the failed drive, was it an external one (USB) or an internal one ?

Also, the pool layout as requested by @Patrick M. Hausen would help...