Read kernel messages with timestamps
Like dmesg, but with real clock time and boot filtering. -k means _TRANSPORT=kernel
journalctl -k -b -o short-precise --grep='segfault|I/O error'
- tool
- journalctl12 commands
- tags
- journalctlkerneldisk
More journalctl commands
-
journalctl --disk-usage; journalctl --vacuum-size=500M --vacuum-time=2dShow the disk usage, then shrink the journal to 500M and delete entries older than 2 days
-
journalctl -t kernel -t sudo -f --no-hostnameFollow by SYSLOG_IDENTIFIER, not by unit. -t can be repeated, --no-hostname removes clutter
-
journalctl -u <name> --output=json --output-fields=__REALTIME_TIMESTAMP,MESSAGE,_PID | jq -r '[.__REALTIME_TIMESTAMP,._PID,.MESSAGE]|@tsv'Keep only the needed JSON fields and print them as tab-separated columns. __REALTIME_TIMESTAMP is in microseconds
-
journalctl -D /mnt/crashed/var/log/journal -b -1 -p errRead journals from another root (rescue or forensics) without copying files
-
journalctl --header | grep -iE 'file|state'; ls -d /var/log/journal 2>/dev/null || echo volatileIf /var/log/journal is missing, logs live in memory and are lost on reboot. Fix: mkdir + systemd-tmpfiles