How to monitor disk activity (I/O) on Linux
How to monitor disk activity (I/O) on Linux?
Use tools like iotop, sar, vmstat, and iostat to discover how busy a disk is and which processes are responsible.
iotopTo find out what processes are actively using the disk, we need to monitor disk I/O. Commands that can be used include iostat, iotop, and vmstat.
Examples using iotop
Run the iotop
command without any parameters to see an interactive display of the total disk read and write speeds, including the processes responsible for initiating these read or write actions.
# iotop
Total DISK READ: 0.00 B/s ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ | Total DISK WRITE: 0.00 B/s
Current DISK READ: 0.00 B/s ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ | Current DISK WRITE: 0.00 B/s
TID PRIO USER DISK READ DISK WRITE GRAPH[R+W]▽ COMMAND [T]
1 be/4 root 0.00 B/s 0.00 B/s systemd ▲
2 be/4 root 0.00 B/s 0.00 B/s kthreadd █
3 be/0 root 0.00 B/s 0.00 B/s rcu_gp █
4 be/0 root 0.00 B/s 0.00 B/s rcu_par_gp █
5 be/0 root 0.00 B/s 0.00 B/s slub_flushwq █
6 be/0 root 0.00 B/s 0.00 B/s netns █
Use ‘q’ to quit.
Non-interactive usage of iotop
Another option is to run iotop
non-interactively. This is achieved by using the --batch option. To make it automatically stop after a few times, we can provide a the --iter=NUMBER option.
# iotop --batch --iter=3
To reduce the output further, we can tell the program to queue the activity by adding a delay with --delay=SECONDS. Instead of a default delay of a single second, we can take a longer period. We need at least two iterations to gather the information. Typically only the active processes are interesting, which we can filter out using the --only option.
# iotop --batch --only --delay=30 --iter=2
Disk activity per device
vmstat
To learn about the most active devices, use the vmstat command. To see a single moment in time, run the command without any parameters.
vmstat
To queue up disk activity and get a better idea on the most active device, let the tool perform two measures with a time interval (e.g. 30 seconds).
# vmstat --disk 30 2
disk- ------------reads------------ ------------writes----------- -----IO------
total merged sectors ms total merged sectors ms cur sec
loop0 276 0 3446 32 0 0 0 0 0 0
loop1 1525 0 131252 192 0 0 0 0 0 4
loop2 361 0 10116 97 0 0 0 0 0 0
loop3 247 0 3484 29 0 0 0 0 0 0
loop4 410 0 10748 141 0 0 0 0 0 0
loop5 404 0 7118 48 0 0 0 0 0 0
loop6 175 0 2276 23 0 0 0 0 0 0
loop7 0 0 0 0 0 0 0 0 0 0
sr0 5 0 2 0 0 0 0 0 0 0
sda 311804 43728 22006273 97600 1922715 958862 116556904 1651642 0 1913
dm-0 341315 0 19843630 120412 2847459 0 113510856 2880140 0 1894
loop0 276 0 3446 32 0 0 0 0 0 0
loop1 1525 0 131252 192 0 0 0 0 0 4
loop2 361 0 10116 97 0 0 0 0 0 0
loop3 247 0 3484 29 0 0 0 0 0 0
loop4 410 0 10748 141 0 0 0 0 0 0
loop5 404 0 7118 48 0 0 0 0 0 0
loop6 175 0 2276 23 0 0 0 0 0 0
loop7 0 0 0 0 0 0 0 0 0 0
sr0 5 0 2 0 0 0 0 0 0 0
sda 311804 43728 22006273 97600 1922737 958875 116557312 1651690 0 1913
dm-0 341315 0 19843630 120412 2847494 0 113511264 2880252 0 1894
iostat
Another tool that provides insights like vmstat
is the the iotop command. By default with colored output and easier to read. Like with other tools we can define a delay and number of measurements.
# iostat 30 2
avg-cpu: %user %nice %system %iowait %steal %idle
0.20 0.01 0.04 0.01 0.01 99.73
Device tps kB_read/s kB_wrtn/s kB_dscd/s kB_read kB_wrtn kB_dscd
dm-0 0.34 1.04 5.92 8.96 9922115 56756396 85913152
loop0 0.00 0.00 0.00 0.00 1723 0 0
loop1 0.00 0.01 0.00 0.00 65626 0 0
loop2 0.00 0.00 0.00 0.00 5058 0 0
loop3 0.00 0.00 0.00 0.00 1742 0 0
loop4 0.00 0.00 0.00 0.00 5374 0 0
loop5 0.00 0.00 0.00 0.00 3559 0 0
loop6 0.00 0.00 0.00 0.00 1138 0 0
sda 0.24 1.15 6.08 9.42 11003436 58279420 90309104
sr0 0.00 0.00 0.00 0.00 1 0 0
avg-cpu: %user %nice %system %iowait %steal %idle
0.43 0.00 0.17 0.10 0.03 99.27
Device tps kB_read/s kB_wrtn/s kB_dscd/s kB_read kB_wrtn kB_dscd
dm-0 1.23 0.00 8.00 0.00 0 240 0
loop0 0.00 0.00 0.00 0.00 0 0 0
loop1 0.00 0.00 0.00 0.00 0 0 0
loop2 0.00 0.00 0.00 0.00 0 0 0
loop3 0.00 0.00 0.00 0.00 0 0 0
loop4 0.00 0.00 0.00 0.00 0 0 0
loop5 0.00 0.00 0.00 0.00 0 0 0
loop6 0.00 0.00 0.00 0.00 0 0 0
sda 0.73 0.00 8.00 0.00 0 240 0
sr0 0.00 0.00 0.00 0.00 0 0 0
When reviewing the data, be aware that the first set of data shows a total number per device. The second one shows the difference, which makes it a great tool to see what happened in the last 30 seconds.