Disk Management

Normal Disk Operations

Add a disk

List the disk currently have in hand. Or just lsblk will also do.

$ lsblk -o NAME,SIZE,FSTYPE,TYPE,MOUNTPOINT
NAME         SIZE FSTYPE            TYPE  MOUNTPOINT
sda          256G                   disk
sdb          256G                   disk /

Format new disks.

$ mkfs -t ext4 /dev/sba
$ mkfs.ext4 /dev/sba

Mount the formatted disk to certain directory.

$ mount /dev/sba /mnt/sba

Add to /etc/fstab so that you don't need to mount it every time the device boots (how does it work?).

$ sudo bash -c 'echo "/dev/sba /mnt/sba ext4 defaults 0 0" >> /etc/fstab

Or use UUID.

$ UUID=$(sudo blkid | grep /dev/sba | cut -f2 -d ' ' | sed -e 's/\"//g')
$ sudo bash -c 'echo "${UUID} /mnt/sba ext4 defaults 0 0" >> /etc/fstab'

Or add a label to the disk and use the label to mount.

$ sudo e2label /dev/sba DISK1
sudo bash -c 'echo "LABEL=DISK1 /mnt/sba ext4 defaults 0 0" >> /etc/fstab'

Optimize disk performance

Adjust the readahead value to increase IO performance

$ sudo blockdev /dev/sba
256

The readahead value is <desired_readahead_bytes> / 512 bytes.

For example, for an 8-MB readahead, 8 MB is 8388608 bytes (8 * 1024 * 1024).

8388608 bytes / 512 bytes = 16384

Set blockdev to 16384 to havea 8-MB readahead.

sudo blockdev --setra 16384 /dev/sba

RAID Operations

The best option to make a software RAID array is mdadm. You can get it from apt or other package manager.

$ sudo apt install mdadm

Normal operations

Check RAID configuration

$ sudo mdadm --detail --scan

Check RAID operation progress / whether there is already a RAID array available

$ cat /proc/mdstat
Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10]
md0 : active raid5 nvme5n1[0] nvme7n1[2] nvme6n1[1] nvme8n1[4] nvme9n1[5]
      60011155456 blocks super 1.2 level 5, 512k chunk, algorithm 2 [5/5] [UUUUU]
      bitmap: 0/112 pages [0KB], 65536KB chunk

unused devices: <none>

Create and resize a RAID array

Create a RAID5 disk array called /dev/md0 with /dev/sda /dev/sdb and /dev/sdc (might take quite some time)

Note that actually a RAID5 disk array can only be named in the form of /dev/md[0-9]+

$ sudo mdadm --create --verbose /dev/md0 \
  --level=5 --raid-devices=3 /dev/sda /dev/sdb /dev/sdc

Grow a RAID5 disk array /dev/md0 with 1 new disk called /dev/sdd (might take quite some time)

$ sudo mdadm --add /dev/md0 /dev/sdd
$ mdadm --grow --raid-devices=5 /dev/md0

Reference

Google Cloud Docs: Optimizing persistent disk performance

Random Notes