nix99 wrote:now I understand if I partition the drive (3tb) in one partition and the partition table is 'loop' will be ok, right?
No, you still don't understand what you've done.
Think of the contents of the disk.
When you start, the disk is blank:
Code: Select all
--------------------------------------------------------------------------------
Here you have run
parted and made a GPT partition. The
GPT below is the partition table:
Code: Select all
-GPT---------------------------------------------------------------------------
Now if you run parted, it will report a "gpt" type of partition table.
Now you have run mkfs.ext4 against the
disk (/dev/sdX) not the
partition (/dev/sdX1), so mkfs has overwritten the previous data with an ext4 filesystem. Now the disk only contains a filesystem:
Code: Select all
ext4ext4ext4ext4ext4ext4ext4ext4ext4ext4ext4ext4ext4ext4ext4ext4ext4ext4ext4---
Now if you run parted, there is no partition table, but parted is smart enough to read the ext4 filesystem and report it is present. However there is still no partition table.
Parted sees there is no partition table so it reports "loop".
There is nothing wrong with this. It will work fine.
However, if you wish to run with a partition table you can, and if you wish to make multiple partitions (multiple filesystems on the one disk) then you
must have a partition table.
So you could make a partition table like before. Here I will use
^ to mark the start of two partitions, though there is no actual mark on the disk:
Code: Select all
-GPT-^---------------------------------------------------^---------------------
Now if you mkfs against the partitions (/dev/sdX1 and /dev/sdX2) you would end up with something like this, where each separate "block" of ext4ext4 is a different filesystem:
Code: Select all
-GPT-^ext4ext4ext4ext4ext4ext4ext4ext4ext4ext4ext4ext4---^ext4ext4ext4ext4ext4-
I hope that makes it a bit clearer.