Just starting off by saying that I am very much an Unraid noob.

I set it up 3 years ago and absolutely love Unraid, but I know that the experts are here on the forum.

---

In my Unraid server, I had 2x 1TB NVMe SSDs which I have configured as a RAID1 pool device.

This stores my appdata, backups, domains and system.

I recently had an alert that 1 of these NVMe drives had failed and then my Docker images and VMs were not responsive.

I rebooted my server and everything was ok.

Then about 1 month later, the other NVMe drive had the same report of failures and I had the same issues.

And again, rebooting resolved the issue.

The odd thing is that neither drive is showing SMART errors, but in the Pool Device Status section, for this pool device, these 2 drives show errors:

This screenshot was taken after I added 2 new NVMe SSDs, but it's the top 2 that were the originals.

My Server

Unraid Version	7.2.0
Processor	Intel 13700k
RAM	64GB DDR5
NVMe SSDs	2x 1TB Samsung 970 Evo Plus (these are the problem drives) 1x 1TB Samsung 9100 Pro 1x 1TB Corsair MP700 Pro

What I Need Help With

I would like to do 2 things:

1. Get this pool device to actually act as a RAID1 and only have 1TB of redundant storage across all 4 NVMe drives.

2. Know if my original 2 NVMe SSDs are failing and need to be replaced as they are still under warranty.

I Know I'm Nuts!!!

I know that a lot of people would say that I'm nuts for having 4 NVMe SSDs in RAID1, but I really don't want my Docker images or VMs going down.

I do back them up every once and a while to an external drive, but I don't want an NVMe failure to cause my server to essentially go down.

I get alerts from my Unraid server whenever there are any issues, but I really want to minimize downtime.

My Overkill to Safeguarding My Data

I overdid things and bought 2 more 1TB NVMe SSDs a few months ago; luckily I bought them before the massive price increases.

Last night I installed those 2 new NVMe SSDS into my system and added them to my RAID1 pool device.

But the odd thing is that this pool device is now showing a size of 2TB rather than the expected 1TB:

This was showing 2TB when I just had the 2 NVMe SSDs in the pool as well.

In the settings for that pool device, I see that the allocation profile is set to raid1:

I backed-up the important contents (appdata & VM images) from this pool device to my main PC and then under the Balance Status section, I selected Convert to raid1 mode from the drop-down and clicked the BALANCE button.

This completed immediately and didn't have any effect.

4 x 1TB drives in raid1= 2 TB usable; that is correct.

Regarding the pool errors, scrub the pool and post the results from the GUI.

42 minutes ago, JorgeB said:

4 x 1TB drives in raid1= 2 TB usable; that is correct.

Regarding the pool errors, scrub the pool and post the results from the GUI.

Thank you for your reply.

I was under the impression that a RAID1 would mirror the data onto all drives in the pool; so that I could loose up-to 3 drives and still not loose any data.

If not, is there a way to get a pool device to function in this way?

Is scrub something that I should be running regularly on a schedule?

Is there a way to and should I also be doing a scrub on my main Unraid array?

I just ran a scrub on this pool and below are the results:

30 minutes ago, MattWeiler said:

I was under the impression that a RAID1 would mirror the data onto all drives in the pool

Nope, RAID 1 means all blocks will exist in two separate devices.

Btrfs also supports 3-way mirror (raid1c3) and 4-way mirror (raid1c4), you can convert using the balance option.

32 minutes ago, MattWeiler said:

Is scrub something that I should be running regularly on a schedule?

Once a month is always good, and every time a pool device drops, like it happened to you, see here for better pool monitoring, at least until 7.3 is out: https://forums.unraid.net/topic/46802-faq-for-unraid-v6/page/2/#findComment-700582

33 minutes ago, MattWeiler said:

Is there a way to and should I also be doing a scrub on my main Unraid array?

In the main array, you typically run a parity check to confirm all is well, but if the file system for the array disks is btrfs or zfs, they can also be scrubbed individually.

17 minutes ago, JorgeB said:

Nope, RAID 1 means all blocks will exist in two separate devices.

Btrfs also supports 3-way mirror (raid1c3) and 4-way mirror (raid1c4), you can convert using the balance option.

Ok, so there is at least a way

Thank you.

15 minutes ago, JorgeB said:

Once a month is always good, and every time a pool device drops, like it happened to you, see here for better pool monitoring, at least until 7.3 is out: https://forums.unraid.net/topic/46802-faq-for-unraid-v6/page/2/#findComment-700582

I'll have my cache pools perform a scrub each month.

15 minutes ago, JorgeB said:

In the main array, you typically run a parity check to confirm all is well, but if the file system for the array disks is btrfs or zfs, they can also be scrubbed individually.

That makes sense.

I have my main array setup to perform a parity check every 3 months.

So far I haven't had any issues with the parity checks.

Thank you for the info, that helps a lot.