1020

I check service status with systemctl status service-name.

By default, I see few rows only, so I add -n50 to see more.

Sometimes, I want to see full log, from start. It could have 1000s of rows.
Now, I check it with -n10000 but that doesn't look like neat solution.

Is there an option to check full systemd service log similar to less command?

0

6 Answers 6

1459

Just use the journalctl command, as in:

journalctl -u service-name.service

Or, to see only log messages for the current boot:

journalctl -u service-name.service -b

For things named <something>.service, you can actually just use <something>, as in:

journalctl -u service-name

But for other sorts of units (sockets, targets, timers, etc), you need to be explicit.

In the above commands, the -u flag is short for --unit, and specifies the name of the unit in which you're interested. -b is short for --boot, and restricts the output to only the current boot so that you don't see lots of older messages. See the journalctl man page for more information.

16
  • 105
    appending --no-pager will print full log, so you wont have to scroll Commented May 29, 2017 at 11:41
  • 127
    appending -e will start the log at the end removing the need to scroll, but without printing the entire log beforehand.
    – timlyo
    Commented Jul 28, 2017 at 14:45
  • 164
    appending -f will follow (print) updates to the log
    – Joe J
    Commented Sep 6, 2017 at 17:10
  • 97
    appending --help will let you see all available options
    – Tzafrir
    Commented Mar 26, 2018 at 20:14
  • 11
    Actually, this was not helpful for me - I want only the output since the unit was (re)started. sudo systemctl -n 99999 status gives me that, but apparently journalctl can only filter "from boot", not "from service start".
    – tobixen
    Commented Jun 26, 2019 at 4:57
80

systemctl can include the complete output of its status listing, without truncation., by adding the -l flag:

systemctl -l status service-name

-l: don't truncate entries with ellipses (...)

--no-pager can be added to avoid invoking a pager when the output is an interactive terminal.

4
  • 7
    Could you explicit your answer somehow? Adding some explanation about the options, and don't hesitate to format your answer!
    – joH1
    Commented Jan 24, 2018 at 10:20
  • 39
    This answer is wrong. The output stays truncated.
    – phil294
    Commented Oct 30, 2018 at 14:21
  • might not be best for current one, but gods... it will make my life easier going forward :D
    – Tom St
    Commented Aug 14, 2021 at 14:51
  • 1
    Since this no longer works on my system, I've posted another answer that makes it posible to get the full logs via systemctl (with full support to journalctl's options to boot). It's a hack but it's been working great for me so far. Commented Oct 17, 2021 at 8:22
44

Use journalctl to View Your System's Logs

          ︙

View journalctl without Paging

To send your logs to standard output and avoid paging them, use the --no-pager option:

journalctl --no-pager

It’s not recommended that you do this without first filtering down the number of logs shown.

journalctl -u service-name.service

          ︙

Show Logs within a Time Range

Use the --since option to show logs after a specified date and time:

journalctl --since "2018-08-30 14:10:10"

Use the --until option to show logs up to a specified date and time:

journalctl --until "2018-09-02 12:05:50"

Combine these to show logs between the two times:

journalctl --since "2018-08-30 14:10:10" --until "2018-09-02 12:05:50"

The above was largely copied from Use journalctl to View Your System's Logs at Linode Docs.  There is more information available there.

6
  • 2
    @RokeJulianLockhart: Please pay more attention to what you are doing.  PagingPermalink and RangePermalink are not code / command / keywords.  champion-runner blindly copied from the source document and somehow converted some non-textual links into the word “Permalink”.  Your edit nearly set his error in concrete. Commented Mar 29 at 20:08
  • In that case, apologies, @G-ManSays'ReinstateMonica'. However, what do those terms refer to, then? That's very unusual standard English, if it is so. Commented Mar 30 at 16:20
  • @RokeJulianLockhart: Apology accepted.   /    Look at the linked page and search for “To send your logs to standard output”.  It’s under a heading, “View journalctl without Paging  #”.  The “#” is a link to “linode.com/docs/guides/how-to-use-journalctl/…”, i.e., a direct link to that section.  All the headings have direct links like that. Commented Mar 30 at 18:16
  • OK, “Permalink” is, obviously, short for “Permanent link”.  It’s a somewhat common web term.  For example, unix.stackexchange.com/a/552945 is a direct link to champion-runner’s answer, which we both edited.  It will take you to the (then) current state of the answer and its comments.  But unix.stackexchange.com/revisions/552945/1 is a permanent link to the first draft/revision of the answer — what it looked like before you edited it.  In other words, a link that is expected to work in the future to get to the exact same place it goes to today. … (Cont’d) Commented Mar 30 at 18:16
  • (Cont’d) …   Now, I don’t understand why linode.com uses the term “permalink” to refer to what I would call a “direct link”, but, apparently, they do.     /     So, to repeat what I said yesterday, champion-runner somehow copied “Paging  #” and pasted “PagingPermalink”. Commented Mar 30 at 18:16
42

Most of the time, it is convenient and easy to use the following bash command:

journalctl -xefu service-name.service

or

journalctl -xefu service-name

It works as if the process is executed via shell and the output is changing dynamically (similar to tail -f).

1
  • 2
    This is the only answer that shows if a unit has been started or stopped (or likewise restart).
    – Dave
    Commented Mar 11, 2022 at 14:19
17

Using journalctl, write logs to a text file, and read it bottom up:

journalctl -u service-name.service > file_name.txt &&\
tail file_name.txt
3
  • Which tail command supports option -r, and what does it do?
    – RalfFriedl
    Commented Nov 19, 2019 at 6:29
  • I like this answer the best. Was able to quickly get real-time log output to a file that I could then winscp to snag it from a Windows box. Commented Sep 9, 2020 at 15:58
  • 1
    If you must use tail like this, at least skip the creation of a file and use pipes instead: journalctl -u service-name.service | tail. Better yet, skip tail and run journalctl -eu service-name.service. Commented Mar 29 at 19:48
7

Since @Julien's answer no longer appears to work on my system (Debian 11), I've finally given in and hijacked systemctl on my system:

systemctl() { 
    if [[ "${1-}" == "log" ]]; then  
        /usr/bin/journalctl -u "${@:2}"; 
    else /usr/bin/systemctl "$@";
    fi 
}

Add this one-liner to your .bashrc and your systemctl will gain a new log "verb" that provides the missing functionality. No more wasting time retyping tedious commands.

As an added bonus, this method will also give you access to all the options of journalctl (provided that they're specified after the unit name):

systemctl log named.service --since=today

Edit: To address John's comment about the lack of bash-completion support and provide a more complete answer, we might as well hijack systemctl's bash completion function too:

source "/usr/share/bash-completion/completions/systemctl"
source "/usr/share/bash-completion/completions/journalctl"

_systemd_log_hijack() {
    if [[ "${COMP_WORDS[1]}" == "log" ]]; then
        if [[ "${COMP_CWORD}" == "2" ]]; then
            COMPREPLY=( $(systemctl list-units 2>&1 | awk -c ' { print $1 }' | grep -F "${COMP_WORDS[2]}") )
        elif [[ "${COMP_CWORD}" -gt "2" ]]; then
            (( COMP_CWORD++ )) || true
            COMP_WORDS=("journalctl" "-u" "${COMP_WORDS[2]}" "${COMP_WORDS[@]:2}")
            COMP_LINE="${COMP_WORDS[*]}"
            _journalctl
        fi  
    else
        _systemctl
    fi

    if [[ "${COMP_CWORD}" == "1" && "log" == "${COMP_WORDS[1]}"* ]]; then
        COMPREPLY=("log" "${COMPREPLY[@]}")
    fi  
}
complete -F _systemd_log_hijack systemctl

This isn't a polished solution but it should provide usable bash completion for the non-standard log command verb.

Keep in mind that this is merely a hack to fix poorly-designed UX. The right way would be to somehow convince the SystemD project to accept a fix upsteam.

3
  • But then it won't show the active/running status since that's not a part of journalctl, right?
    – Kevin
    Commented Jun 18 at 18:15
  • the time gained by this alias is then lost by the breakage of bash-completion Commented Jul 7 at 5:14
  • @JohanBoulé Very true but slightly mitigable with yet another hijack. Commented Jul 7 at 9:29

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .