raid5 with 3 hard drives (12To) create 4 partitions...!?

Hello,

I never see this before. We use 3 hard drives (12To each) to build a raid5 system.

I created a raid 5 like usual, but OMV create not only 1 partition but 4 partitions. The normal partition have 21.83Tb with state clean, and there are 3 smaller partitions (867GB, 811GB, 915GB) with state = false

So I read doc, and search in the forum, wait until the raid is completely build, reboot, etc but the 3 small partitions are still there. See my screenshot for more understanding

I don't understand.

If someone have a idea to fix the problem...??

Thanh you.

Regard.

Quote from ryecoaaron

OMV won't create any partitions for a raid array. The partitions must have already been on the drives. I would wipe them and recreate (probably wouldn't use raid either).

I already do a wipe (quick) for all the 3 disks
If a delete one of the strange device like md0p3, all devices are deleted

So now I do a secure wipe for one disk, just to be sure... It takes a lot of time.

And I can see on my last screenshot that the problem is not sdb, sdc or sdd but the raid himself.
What are the origines of the md0p2, md0p3 and md0p4 ?

Maybe the system doesn't support 12To disk capacity..?

And for info, there is no plugins installed.

Thank for any help.

Regard.
Piet.

Hello,

ok I delete mbr for sdb, sdca and sdd with dd if=/dev/zero of=/dev/sd*extension disk* bs=512 count=1
Nothing changes

I also wipe (quick) again the three disk. Nothing changes.

And I wipe (secure) the disk sdd, nothing changes.

Ah yes 2 things changes, the capacity of md0p1, md0p3 and md0p4

And there is no md0p2 now, it is replaced by md0p1...!

See difference between my first screenshot.

Regard.
Piet

Hello,

ok I do

wipefs -a /dev/sdb
wipefs -a /dev/sdc
wipefs -a /dev/sdd

(see screenshot)

And create again a raid 5.
And NO, it is the same problem.. (see my screenshot)

Maybe the problem are the big disks? Too big (12To) ?

Regard.
Piet

Quote from tkaiser

For RAID5 they're way too big unless you got some magic drives with an unbelievably amazing unrecoverable read error rate (URE). See the first answer you got in this thread.

This is not the problem. This is just my choice if I want to made a raid 5 with 3 12To disks.
The problem are the 3 stranges devices created. Strange ins't it ?

If someone have a idea to fix it ? And I suppose that OMV support 12To disk..?

Regard.
Piet

Quote from cabrio_leo

Here is an interesting ZDNet article about RAID 5 from 2017 and why it can be sometimes more efficient to restore a backup instead of rebuilding a failed drive: Why RAID 5 still works, usuallyAnd another one: Why the best RAID configuration is no RAID configuration

Don't understand. I read the article, but can't see any explication about the problem here. The problem is not recontruction, it is a creation.

So if someone have a idea for the problem ?

Regard.
Piet

We have a first backup (Nakivo on Synology). This is for the second backup. And, no offense, but you are off topic. Do you have a real idea to fix the problem ?

Or anybody have a idea for this problem ?

Thank you.

Regard.

Piet