1
1
mirror of https://github.com/systemd/systemd-stable.git synced 2024-12-23 17:34:00 +03:00
systemd-stable/man/systemd-journald.service.xml
Lennart Poettering ec20fe5ffb journald: make maximum size of stream log lines configurable and bump it to 48K (#6838)
This adds a new setting LineMax= to journald.conf, and sets it by
default to 48K. When we convert stream-based stdout/stderr logging into
record-based log entries, read up to the specified amount of bytes
before forcing a line-break.

This also makes three related changes:

- When a NUL byte is read we'll not recognize this as alternative line
  break, instead of silently dropping everything after it. (see #4863)

- The reason for a line-break is now encoded in the log record, if it
  wasn't a plain newline. Specifically, we distuingish "nul",
  "line-max" and "eof", for line breaks due to NUL byte, due to the
  maximum line length as configured with LineMax= or due to end of
  stream. This data is stored in the new implicit _LINE_BREAK= field.
  It's not synthesized for plain \n line breaks.

- A randomized 128bit ID is assigned to each log stream.

With these three changes in place it's (mostly) possible to reconstruct
the original byte streams from log data, as (most) of the context of
the conversion from the byte stream to log records is saved now. (So,
the only bits we still drop are empty lines. Which might be something to
look into in a future change, and which is outside of the scope of this
work)

Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=86465
See: #4863
Replaces: #4875
2017-09-22 10:22:24 +02:00

320 lines
15 KiB
XML

<?xml version='1.0'?> <!--*-nxml-*-->
<!DOCTYPE refentry PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN"
"http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd">
<!--
This file is part of systemd.
Copyright 2010 Lennart Poettering
systemd is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License as published by
the Free Software Foundation; either version 2.1 of the License, or
(at your option) any later version.
systemd is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public License
along with systemd; If not, see <http://www.gnu.org/licenses/>.
-->
<refentry id="systemd-journald.service">
<refentryinfo>
<title>systemd-journald.service</title>
<productname>systemd</productname>
<authorgroup>
<author>
<contrib>Developer</contrib>
<firstname>Lennart</firstname>
<surname>Poettering</surname>
<email>lennart@poettering.net</email>
</author>
</authorgroup>
</refentryinfo>
<refmeta>
<refentrytitle>systemd-journald.service</refentrytitle>
<manvolnum>8</manvolnum>
</refmeta>
<refnamediv>
<refname>systemd-journald.service</refname>
<refname>systemd-journald.socket</refname>
<refname>systemd-journald-dev-log.socket</refname>
<refname>systemd-journald-audit.socket</refname>
<refname>systemd-journald</refname>
<refpurpose>Journal service</refpurpose>
</refnamediv>
<refsynopsisdiv>
<para><filename>systemd-journald.service</filename></para>
<para><filename>systemd-journald.socket</filename></para>
<para><filename>systemd-journald-dev-log.socket</filename></para>
<para><filename>systemd-journald-audit.socket</filename></para>
<para><filename>/usr/lib/systemd/systemd-journald</filename></para>
</refsynopsisdiv>
<refsect1>
<title>Description</title>
<para><filename>systemd-journald</filename> is a system service
that collects and stores logging data. It creates and maintains
structured, indexed journals based on logging information that is
received from a variety of sources:</para>
<itemizedlist>
<listitem><para>Kernel log messages, via kmsg</para></listitem>
<listitem><para>Simple system log messages, via the <filename>libc</filename> <citerefentry
project='man-pages'><refentrytitle>syslog</refentrytitle><manvolnum>3</manvolnum></citerefentry>
call</para></listitem>
<listitem><para>Structured system log messages via the native
Journal API, see
<citerefentry><refentrytitle>sd_journal_print</refentrytitle><manvolnum>4</manvolnum></citerefentry></para></listitem>
<listitem><para>Standard output and standard error of service units. For further details see
below.</para></listitem>
<listitem><para>Audit records, originating from the kernel audit subsystem</para></listitem>
</itemizedlist>
<para>The daemon will implicitly collect numerous metadata fields
for each log messages in a secure and unfakeable way. See
<citerefentry><refentrytitle>systemd.journal-fields</refentrytitle><manvolnum>7</manvolnum></citerefentry>
for more information about the collected metadata.
</para>
<para>Log data collected by the journal is primarily text-based
but can also include binary data where necessary. All objects
stored in the journal can be up to 2^64-1 bytes in size.</para>
<para>By default, the journal stores log data in
<filename>/run/log/journal/</filename>. Since
<filename>/run/</filename> is volatile, log data is lost at
reboot. To make the data persistent, it is sufficient to create
<filename>/var/log/journal/</filename> where
<filename>systemd-journald</filename> will then store the
data:</para>
<programlisting>mkdir -p /var/log/journal
systemd-tmpfiles --create --prefix /var/log/journal</programlisting>
<para>See
<citerefentry><refentrytitle>journald.conf</refentrytitle><manvolnum>5</manvolnum></citerefentry>
for information about the configuration of this service.</para>
</refsect1>
<refsect1>
<title>Stream logging</title>
<para>The systemd service manager invokes all service processes with standard output and standard error connected
to the journal by default. This behaviour may be altered via the
<varname>StandardOutput=</varname>/<varname>StandardError=</varname> unit file settings, see
<citerefentry><refentrytitle>systemd.exec</refentrytitle><manvolnum>5</manvolnum></citerefentry> for details. The
journal converts the log byte stream received this way into individual log records, splitting the stream at newline
(<literal>\n</literal>, ASCII <constant>10</constant>) and <constant>NUL</constant> bytes.</para>
<para>If <filename>systemd-journald.service</filename> is stopped, the stream connections associated with all
services are terminated. Further writes to those streams by the service will result in <constant>EPIPE</constant>
errors. In order to react gracefully in this case it is recommended that programs logging to standard output/error
ignore such errors. If the the <constant>SIGPIPE</constant> UNIX signal handler is not blocked or turned off, such
write attempts will also result in such process signals being generated, see
<citerefentry><refentrytitle>signal</refentrytitle><manvolnum>7</manvolnum></citerefentry>. To mitigate this issue,
systemd service manager explicitly turns off the <constant>SIGPIPE</constant> signal for all invoked processes by
default (this may be changed for each unit individually via the <varname>IgnoreSIGPIPE=</varname> option, see
<citerefentry><refentrytitle>systemd.exec</refentrytitle><manvolnum>5</manvolnum></citerefentry> for
details). After the standard output/standard error streams have been terminated they may not be recovered until the
services they are associated with are restarted. Note that during normal operation,
<filename>systemd-journald.service</filename> stores copies of the file descriptors for those streams in the
service manager. If <filename>systemd-journald.service</filename> is restarted using <command>systemctl
restart</command> or equivalent operation instead of a pair of separate <command>systemctl stop</command> and
<command>systemctl start</command> commands (or equivalent operations), these stream connections are not terminated
and survive the restart. It is thus safe to restart <filename>systemd-journald.service</filename>, but stopping it
is not recommended.</para>
<para>Note that the log record metadata for records transferred via such standard output/error streams reflect the
metadata of the peer the stream was originally created for. If the stream connection is passed on to other
processes (such as further child processes forked off the main service process), the log records will not reflect
their metadata, but will continue to describe the original process. This is different from the other logging
transports listed above, which are inherently record based and where the metadata is always associated with the
individual record.</para>
<para>In addition to the the implicit standard output/error logging of services, stream logging is also available
via the <citerefentry><refentrytitle>systemd-cat</refentrytitle><manvolnum>1</manvolnum></citerefentry> command
line tool.</para>
<para>Currently, the number of parallel log streams <filename>systemd-journald</filename> will accept is limited to
4096. When this limit is reached further log streams may be established but will receieve
<constant>EPIPE</constant> right from the beginning.</para>
</refsect1>
<refsect1>
<title>Signals</title>
<variablelist>
<varlistentry>
<term>SIGUSR1</term>
<listitem><para>Request that journal data from
<filename>/run/</filename> is flushed to
<filename>/var/</filename> in order to make it persistent (if
this is enabled). This must be used after
<filename>/var/</filename> is mounted, as otherwise log data
from <filename>/run</filename> is never flushed to
<filename>/var</filename> regardless of the configuration. The
<command>journalctl --flush</command> command uses this signal
to request flushing of the journal files, and then waits for
the operation to complete. See
<citerefentry><refentrytitle>journalctl</refentrytitle><manvolnum>1</manvolnum></citerefentry>
for details.</para></listitem>
</varlistentry>
<varlistentry>
<term>SIGUSR2</term>
<listitem><para>Request immediate rotation of the journal
files. The <command>journalctl --rotate</command> command uses
this signal to request journal file
rotation.</para></listitem>
</varlistentry>
<varlistentry>
<term>SIGRTMIN+1</term>
<listitem><para>Request that all unwritten log data is written
to disk. The <command>journalctl --sync</command> command uses
this signal to trigger journal synchronization, and then waits
for the operation to complete.</para></listitem>
</varlistentry>
</variablelist>
</refsect1>
<refsect1>
<title>Kernel Command Line</title>
<para>A few configuration parameters from
<filename>journald.conf</filename> may be overridden on the kernel
command line:</para>
<variablelist class='kernel-commandline-options'>
<varlistentry>
<term><varname>systemd.journald.forward_to_syslog=</varname></term>
<term><varname>systemd.journald.forward_to_kmsg=</varname></term>
<term><varname>systemd.journald.forward_to_console=</varname></term>
<term><varname>systemd.journald.forward_to_wall=</varname></term>
<listitem><para>Enables/disables forwarding of collected log
messages to syslog, the kernel log buffer, the system console
or wall.
</para>
<para>See
<citerefentry><refentrytitle>journald.conf</refentrytitle><manvolnum>5</manvolnum></citerefentry>
for information about these settings.</para>
</listitem>
</varlistentry>
</variablelist>
</refsect1>
<refsect1>
<title>Access Control</title>
<para>Journal files are, by default, owned and readable by the
<literal>systemd-journal</literal> system group but are not
writable. Adding a user to this group thus enables her/him to read
the journal files.</para>
<para>By default, each logged in user will get her/his own set of
journal files in <filename>/var/log/journal/</filename>. These
files will not be owned by the user, however, in order to avoid
that the user can write to them directly. Instead, file system
ACLs are used to ensure the user gets read access only.</para>
<para>Additional users and groups may be granted access to journal
files via file system access control lists (ACL). Distributions
and administrators may choose to grant read access to all members
of the <literal>wheel</literal> and <literal>adm</literal> system
groups with a command such as the following:</para>
<programlisting># setfacl -Rnm g:wheel:rx,d:g:wheel:rx,g:adm:rx,d:g:adm:rx /var/log/journal/</programlisting>
<para>Note that this command will update the ACLs both for
existing journal files and for future journal files created in the
<filename>/var/log/journal/</filename> directory.</para>
</refsect1>
<refsect1>
<title>Files</title>
<variablelist>
<varlistentry>
<term><filename>/etc/systemd/journald.conf</filename></term>
<listitem><para>Configure
<command>systemd-journald</command>
behavior. See
<citerefentry><refentrytitle>journald.conf</refentrytitle><manvolnum>5</manvolnum></citerefentry>.
</para></listitem>
</varlistentry>
<varlistentry>
<term><filename>/run/log/journal/<replaceable>machine-id</replaceable>/*.journal</filename></term>
<term><filename>/run/log/journal/<replaceable>machine-id</replaceable>/*.journal~</filename></term>
<term><filename>/var/log/journal/<replaceable>machine-id</replaceable>/*.journal</filename></term>
<term><filename>/var/log/journal/<replaceable>machine-id</replaceable>/*.journal~</filename></term>
<listitem><para><command>systemd-journald</command> writes
entries to files in
<filename>/run/log/journal/<replaceable>machine-id</replaceable>/</filename>
or
<filename>/var/log/journal/<replaceable>machine-id</replaceable>/</filename>
with the <literal>.journal</literal> suffix. If the daemon is
stopped uncleanly, or if the files are found to be corrupted,
they are renamed using the <literal>.journal~</literal>
suffix, and <command>systemd-journald</command> starts writing
to a new file. <filename>/run</filename> is used when
<filename>/var/log/journal</filename> is not available, or
when <option>Storage=volatile</option> is set in the
<citerefentry><refentrytitle>journald.conf</refentrytitle><manvolnum>5</manvolnum></citerefentry>
configuration file.</para></listitem>
</varlistentry>
<varlistentry>
<term><filename>/dev/kmsg</filename></term>
<term><filename>/dev/log</filename></term>
<term><filename>/run/systemd/journal/dev-log</filename></term>
<term><filename>/run/systemd/journal/socket</filename></term>
<term><filename>/run/systemd/journal/stdout</filename></term>
<listitem><para>Sockets and other paths that
<command>systemd-journald</command> will listen on that are
visible in the file system. In addition to these, journald can
listen for audit events using netlink.</para></listitem>
</varlistentry>
</variablelist>
</refsect1>
<refsect1>
<title>See Also</title>
<para>
<citerefentry><refentrytitle>systemd</refentrytitle><manvolnum>1</manvolnum></citerefentry>,
<citerefentry><refentrytitle>journalctl</refentrytitle><manvolnum>1</manvolnum></citerefentry>,
<citerefentry><refentrytitle>journald.conf</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
<citerefentry><refentrytitle>systemd.journal-fields</refentrytitle><manvolnum>7</manvolnum></citerefentry>,
<citerefentry><refentrytitle>sd-journal</refentrytitle><manvolnum>3</manvolnum></citerefentry>,
<citerefentry><refentrytitle>systemd-coredump</refentrytitle><manvolnum>8</manvolnum></citerefentry>,
<citerefentry project='die-net'><refentrytitle>setfacl</refentrytitle><manvolnum>1</manvolnum></citerefentry>,
<citerefentry><refentrytitle>sd_journal_print</refentrytitle><manvolnum>4</manvolnum></citerefentry>,
<command>pydoc systemd.journal</command>
</para>
</refsect1>
</refentry>