Logic's Site

Appearance

IO Tools ​

Disk Usage ​

df && du ​

bash
# filesystem-level free space
df -h
df -hT                                   # include filesystem type
df -ih                                   # inode usage (often what runs out first)

# directory-level usage
du -sh /var/log                          # summary, human-readable
du -sh /var/* | sort -h                  # top-level breakdown sorted
du -h --max-depth=1 /var | sort -h
du -ah /var | sort -h | tail -20         # top 20 largest files/dirs

ncdu ​

bash
# interactive disk usage analyzer (apt install ncdu)
ncdu /
ncdu -x /                                # stay on one filesystem

lsof (Disk Hogs) ​

bash
# find what is holding a deleted file (common cause of "df says full, du says not")
lsof +L1 | grep -v "^COMMAND"

# files open under a path
lsof +D /var/log

Block Devices ​

blkid && lsblk ​

bash
# show all block device info
blkid
lsblk -f /dev/sda
lsblk -o UUID,PARTUUID,PATH,MOUNTPOINT /dev/sdb
lsblk -d -o NAME,ROTA,SIZE,MODEL          # ROTA=0 for SSD, 1 for HDD

SCSI Bus Rescan ​

bash
# detect newly added disks without rebooting
for host in $(ls /sys/class/scsi_host); do
  echo "- - -" > /sys/class/scsi_host/$host/scan
done

Partitioning ​

bash
# MBR: fdisk, parted
# GPT: gdisk, parted

# show disk partition info
fdisk -l /dev/sda
gdisk -l /dev/sda
parted -l /dev/sda

fdisk (MBR) ​

bash
fdisk /dev/sda

gdisk (GPT) ​

bash
gdisk /dev/sda

parted ​

bash
# MBR (Legacy Boot) layout
parted /dev/sda -- mklabel msdos
parted /dev/sda -- mkpart primary 1MB -2GB
parted /dev/sda -- set 1 boot on
parted /dev/sda -- mkpart primary linux-swap -2GB 100%

# GPT (UEFI) layout
parted /dev/sdb -- unit mib
parted /dev/sdb -- mklabel gpt
parted /dev/sdb -- mkpart primary 1 3
parted /dev/sdb -- mkpart ESP fat32 3 515
parted /dev/sdb -- mkpart root ext4 515 -1
parted /dev/sdb -- set 1 bios_grub on
parted /dev/sdb -- set 2 esp on

partprobe && growpart ​

bash
# reread partition table without reboot
partprobe

# grow a partition to fill available space (cloud-utils-growpart)
growpart /dev/sda 1

Filesystem ​

mkfs ​

bash
# ext4
mkfs.ext4 /dev/sda1

# xfs
mkfs.xfs /dev/sda1

# mount and persist
mount /dev/sda1 /data
echo 'UUID=<uuid> /data ext4 defaults 0 2' >> /etc/fstab

Resize ​

bash
# resize after growpart (ext2/3/4)
resize2fs /data

# resize xfs (must be mounted)
xfs_growfs -d /data

tune2fs && xfs_info ​

bash
# ext4: show / tune filesystem parameters
tune2fs -l /dev/sda1
tune2fs -m 1 /dev/sda1                   # reduce reserved-for-root from 5% to 1%
tune2fs -c 0 -i 0 /dev/sda1              # disable fsck mount-count / time check

# xfs: show geometry
xfs_info /data

Mount & Swap ​

mount && findmnt ​

bash
# show mounted filesystems
mount | column -t
findmnt
findmnt -T /var/log                      # which mount serves a path
findmnt --verify                         # validate /etc/fstab

# mount / remount
mount /dev/sda1 /data
mount -o remount,rw /                    # flip readonly root to rw

Swap ​

bash
# show swap usage
swapon --show
cat /proc/swaps
free -h

# enable / disable
swapon /swapfile
swapoff -a

# create a swap file
dd if=/dev/zero of=/swapfile bs=1M count=2048
chmod 600 /swapfile
mkswap /swapfile
swapon /swapfile

I/O Benchmark ​

dd ​

bash
# CPU baseline (no disk involved)
dd if=/dev/zero of=/dev/null

# sequential write throughput (direct I/O, bypass page cache)
time dd if=/dev/zero of=test.file bs=1G count=2 oflag=direct

fio ​

bash
# sequential read
fio -filename=/tmp/test.file -direct=1 -iodepth 1 -thread -rw=read \
    -ioengine=psync -bs=16k -size=2G -numjobs=10 -runtime=60 \
    -group_reporting -name=test_r

# sequential write
fio -filename=/tmp/test.file -direct=1 -iodepth 1 -thread -rw=write \
    -ioengine=psync -bs=16k -size=2G -numjobs=10 -runtime=60 \
    -group_reporting -name=test_w

# random write
fio -filename=/tmp/test.file -direct=1 -iodepth 1 -thread -rw=randwrite \
    -ioengine=psync -bs=16k -size=2G -numjobs=10 -runtime=60 \
    -group_reporting -name=test_randw

# mixed random read/write (70% read, noop scheduler)
fio -filename=/var/test.file -direct=1 -iodepth 1 -thread -rw=randrw -rwmixread=70 \
    -ioengine=psync -bs=16k -size=2G -numjobs=10 -runtime=60 \
    -group_reporting -name=test_r_w -ioscheduler=noop

I/O Monitoring ​

iostat ​

bash
# install
apt install sysstat

# usage
iostat [options] [delay [count]]
iostat 2                                 # uninterrupted, every 2s
iostat 1 10                              # 10 samples, 1s interval

# common flags
-c     CPU utilization
-d     device utilization
-h     human-readable
-x     extended statistics
-t     timestamp

# example: extended stats for sda and sdb, every 1s, 10 samples
iostat -dhx sda sdb 1 10

iotop ​

bash
# common flags
-o, --only            only show processes/threads doing I/O
-b, --batch           non-interactive mode
-n NUM                number of iterations
-d SEC                delay between iterations [1s]
-p PID                processes to monitor
-u USER               users to monitor
-P                    only processes (not threads)
-a, --accumulated     accumulated I/O instead of bandwidth
-k                    kilobytes instead of human-friendly
-t                    timestamp per line (implies --batch)

# accumulated, process-level, only active I/O
iotop -oPa

pidstat ​

bash
# per-process I/O, 1s interval
pidstat -d 1

# per-process CPU and I/O, 5 samples
pidstat -du 1 5

sar (Disk I/O) ​

bash
# overall I/O transfer rate
sar -b 1 10

# per-block-device activity (pretty-print names)
sar -d -p 1 10

vmstat (Disk) ​

bash
# disk statistics with timestamps
vmstat -d -t

# event counters and memory stats
vmstat -s

# number of forks since boot
vmstat -f

Syscall & File Tracing ​

strace ​

bash
# install
apt install strace

# top-of-time syscall summary (great first look)
strace -c ls

# trace only file-related syscalls
strace -e trace=file -p <PID>
strace -e trace=openat,read,write -p <PID>

# follow forks
strace -f -p <PID>

# write trace to file
strace -o /tmp/strace.out -f -tt -p <PID>

lsof (Open Files) ​

bash
# all open files for a process
lsof -p <PID>

# who has this file open
lsof /var/log/syslog

# files opened by a user
lsof -u username

# all open files of a process tree
lsof -R -p <PID>

Reference:

  1. fio Documentation
  2. sysstat (iostat/pidstat/sar)
  3. iotop
  4. util-linux (lsblk/blkid/fdisk)
  5. GNU parted
  6. strace
  7. lsof

Power by VitePress & Vue

Copyright © 2025-present Logic