Linux Command Reference ยท Packages and Services
Linux dmesg command
Reference for the Linux dmesg command: read kernel messages from the ring buffer, filter by level, show readable timestamps, follow new output and fix permission errors.
dmesg prints the kernel's ring buffer: a fixed-size memory area where the kernel stores its own log messages. It records what happens at boot (hardware detection, drivers, filesystems mounted) and later events such as USB devices plugged in, disk errors, network link changes, and the out-of-memory killer ending a process.
When something hardware- or driver-related goes wrong, dmesg is one of the first places to look. The buffer has a limited size, so the oldest lines are overwritten on a busy system; for permanent history use journalctl -k.
dmesg | tail -20 sudo dmesg --level=err,warn dmesg -T | grep -i usb
Try it
Reading the kernel buffer is often restricted in containers and sandboxes, so the first example reports what happens here. The second example filters sample kernel lines with the same patterns you would use on a real system.
Press Run (or Ctrl + Enter) to see the result here.
Program input ยท what you'd type while the program runs
Type the answers before you press Run, one per line. Each time the program reads a line (with read or cat), it gets the next line from this box.
Press Run (or Ctrl + Enter) to see the result here.
Program input ยท what you'd type while the program runs
Type the answers before you press Run, one per line. Each time the program reads a line (with read or cat), it gets the next line from this box.
Syntax
dmesg [OPTION]...
| Option | Description |
|---|---|
-T, --ctime |
Human-readable timestamps (approximate: based on boot time). |
-H, --human |
Readable output with colors, relative times and a pager. |
-l LIST, --level=LIST |
Only these levels: emerg, alert, crit, err, warn, notice, info, debug. |
-f LIST, --facility=LIST |
Only these facilities (kern, user, daemon, ...). |
-w, --follow |
Keep running and print new messages as they appear. |
-W, --follow-new |
Like -w but skip existing lines. |
-c, --read-clear |
Print, then clear the buffer (root). |
-C, --clear |
Clear without printing (root). |
-n LEVEL |
Set the console log level (root). |
-x, --decode |
Show facility and level as text. |
-k |
Kernel messages only. |
Returns: 0 on success; non-zero if the buffer cannot be read, for example "Operation not permitted".
Usage notes
- Permission denied: many distributions set
kernel.dmesg_restrict=1, so ordinary users needsudoor membership of a log group. Check withsysctl kernel.dmesg_restrict. - Timestamps are seconds since boot.
-Tconverts them to wall-clock time but can be wrong after suspend or long uptime, because the clock the kernel counts with pauses during sleep. - The buffer is finite (set at boot, often 256 KiB or more). After a crash or reboot it is gone;
journalctl -k -b -1shows the previous boot if the journal is persistent. - Typical searches:
dmesg | grep -i -E 'oom|killed process'for memory kills;grep -i 'I/O error'for failing disks;grep -i firmwarefor missing driver files. - Following events: run
dmesg -win one terminal, then plug in a device in another; the new lines tell you which device name it received. - Clearing the buffer hides evidence. Avoid
-con a machine you are troubleshooting. - Containers usually share the host's kernel, but container runtimes often block
dmesgfor safety. - Levels are numbers 0 (emerg) to 7 (debug); only messages at or above the console level are printed to the screen, but all of them are stored in the buffer.
Related
journalctl,uname,logger,lsmod and modprobe- Lessons: Boot and kernel, Disks and logs