More flexible Storage Pools and Volumes options

[AlAleksiev]

In TOS 6.x we have 2 disks that need to be selected for the main system (core and swap) - those are using mandatory RAID1.
Usually we have identical disks in the NAS bays - to provide easy replacement in case of failure.
Let's take the simple case with 5 disks - the space will look like that:


In the above is visible that we have 10GB wasted space for every 3rd, 4th etc. disk.
The ldblk will show that clearly on the CLI:

Code: Select all

$ lsblk
NAME          MAJ:MIN RM  SIZE RO TYPE  MOUNTPOINTS
sda             8:0    0  3.6T  0 disk
├─sda1          8:1    0  285M  0 part
├─sda2          8:2    0  7.6G  0 part
│ └─md9         9:9    0  7.6G  0 raid1 /
├─sda3          8:3    0  1.9G  0 part
│ └─md8         9:8    0  1.9G  0 raid1 [SWAP]
└─sda4          8:4    0  3.6T  0 part
  └─md0         9:0    0 10.9T  0 raid6
    ├─vg0-lv0 252:0    0    9T  0 lvm   /Volume1
    │                                   /var/subvols/8vEbTxkKvwa
    └─vg0-lv1 252:1    0  1.9T  0 lvm   /Volume2
                                        /var/subvols/8vEbTxkKvwb
sdb             8:16   0  3.6T  0 disk
├─sdb1          8:17   0  285M  0 part
├─sdb2          8:18   0  7.6G  0 part
│ └─md9         9:9    0  7.6G  0 raid1 /
├─sdb3          8:19   0  1.9G  0 part
│ └─md8         9:8    0  1.9G  0 raid1 [SWAP]
└─sdb4          8:20   0  3.6T  0 part
  └─md0         9:0    0 10.9T  0 raid6
    ├─vg0-lv0 252:0    0    9T  0 lvm   /Volume1
    │                                   /var/subvols/8vEbTxkKvwa
    └─vg0-lv1 252:1    0  1.9T  0 lvm   /Volume2
                                        /var/subvols/8vEbTxkKvwb
sdc             8:32   0  3.6T  0 disk
├─sdc1          8:33   0  285M  0 part    <---------- waste
├─sdc2          8:34   0  7.6G  0 part    <---------- waste
├─sdc3          8:35   0  1.9G  0 part    <---------- waste
└─sdc4          8:36   0  3.6T  0 part
  └─md0         9:0    0 10.9T  0 raid6
    ├─vg0-lv0 252:0    0    9T  0 lvm   /Volume1
    │                                   /var/subvols/8vEbTxkKvwa
    └─vg0-lv1 252:1    0  1.9T  0 lvm   /Volume2
                                        /var/subvols/8vEbTxkKvwb
sdd             8:48   0  3.6T  0 disk
├─sdd1          8:49   0  285M  0 part    <---------- waste
├─sdd2          8:50   0  7.6G  0 part    <---------- waste
├─sdd3          8:51   0  1.9G  0 part    <---------- waste
└─sdd4          8:52   0  3.6T  0 part
  └─md0         9:0    0 10.9T  0 raid6
    ├─vg0-lv0 252:0    0    9T  0 lvm   /Volume1
    │                                   /var/subvols/8vEbTxkKvwa
    └─vg0-lv1 252:1    0  1.9T  0 lvm   /Volume2
                                        /var/subvols/8vEbTxkKvwb
sde             8:64   0  3.6T  0 disk
├─sde1          8:65   0  285M  0 part    <---------- waste
├─sde2          8:66   0  7.6G  0 part    <---------- waste
├─sde3          8:67   0  1.9G  0 part    <---------- waste
└─sde4          8:68   0  3.6T  0 part
  └─md0         9:0    0 10.9T  0 raid6
    ├─vg0-lv0 252:0    0    9T  0 lvm   /Volume1
    │                                   /var/subvols/8vEbTxkKvwa
    └─vg0-lv1 252:1    0  1.9T  0 lvm   /Volume2
                                        /var/subvols/8vEbTxkKvwb

We can manually provision that space, with RAID5 for example, like that:

Code: Select all

mdadm --create /dev/md10 --level=5 --raid-devices=3 /dev/sdc2 /dev/sdd2 /dev/sde2
mdadm --create /dev/md11 --level=5 --raid-devices=3 /dev/sdc3 /dev/sdd3 /dev/sde3
cat /proc/mdstat
mdadm -E /dev/sd[c-e][2-3]
mdadm --detail /dev/md10
mdadm --detail /dev/md11

pvcreate /dev/md10 /dev/md11
lvmdiskscan -l
vgcreate vg10 /dev/md10 /dev/md11
lvcreate -n lv10 -l 100%FREE vg10
lvdisplay
mkfs.btrfs /dev/vg10/lv10

mkdir /MyVolume
mount /dev/vg10/lv10 /MyVolume

It's not going to be so easy to keep that config permanent (to survive the reload) by two reasons:
1. The files /etc/mdadm/mdadm.conf and /etc/mdadm.conf are rewritten during the reload
2. The command "update-initramfs -u" is not available

For 5 disks the RAID5 will give us protected 20GB, for 8 disks that will be 50GB etc.
It will be great if the TOS 6.x (and later 7.x) is capable to provision those spaces natively, isn't it?

[CursaYang]

Starting from the TOS 6.0.7xx version, we have added a system disk management feature, allowing you to choose any hard drive for system installation.

Regarding the reserved space issue you mentioned, when creating a storage pool with hard disks in NAS, system partitions and swap partitions are reserved. This allows customers to transfer system space at any time without affecting the data partitions on the hard disk.

[AlAleksiev]

I can see the idea... thank you for the clarification!
It's not up to the user as a choice, but I got it.

What will happen if few disks are bigger than the two used for the system space - they'll have extra room for additional provisioning.
Something like that:

Code: Select all

Disk1[4TB]   System(10GB)  TRAID(3.6TB)
Disk2[4TB]   System(10GB)  TRAID(3.6TB)
DIsk3[4TB] reserved(10GB)  TRAID(3.6TB)
Disk4[6TB] reserved(10GB)  TRAID(3.6TB)  residual(2TB)
Disk5[6TB] reserved(10GB)  TRAID(3.6TB)  residual(2TB)
Disk6[6TB] reserved(10GB)  TRAID(3.6TB)  residual(2TB)

TOS 6.0.7XX and 7.X will allow us to provision those 3 x 2TB, let say with RAID5?

[CursaYang]

When you choose to create a TRAID mode storage pool, it will automatically use the extra space to create an array.
For specific details, please refer to the official website introduction: https://www.terra-master.com/zh-cn/page ... ster-traid

[AlAleksiev]

I agree - TRAID+ is introducing some benefits. I'm not questioning them - just looking for more flexibility for the users.
That picture should give you idea what I mean:


Cheers!

[CursaYang]

Defining redundant space flexibly can indeed provide more usable space and reduce waste. However, not all customers have your level of expertise; in most cases, they are more likely to use hard drives of the same capacity to set up RAID. Nevertheless, this is still a good suggestion, and I will feed this requirement back to the product team for evaluation.

[AlAleksiev]

Thank you!