This article describes the command-line tools and its options in fluentd project.
fluentd
Following are the fluentd command-line options (fluentd -h):
Usage: fluentd [options]
-s, --setup [DIR=/etc/fluent] install sample configuration file to the directory
-c, --config PATH config file path (default: /etc/fluent/fluent.conf)
--dry-run Check fluentd setup is correct or not
--show-plugin-config=PLUGIN [DEPRECATED] Show PLUGIN configuration and exit(ex: input:dummy)
-p, --plugin DIR add plugin directory
-I PATH add library path
-r NAME load library
-d, --daemon PIDFILE daemonize fluent process
--under-supervisor run fluent worker under supervisor (this option is NOT for users)
--no-supervisor run fluent worker without supervisor
--workers NUM specify the number of workers under supervisor
--user USER change user
--group GROUP change group
--umask UMASK change umask
-o, --log PATH log file path
--log-rotate-age AGE generations to keep rotated log files
--log-rotate-size BYTES sets the byte size to rotate log files
--log-event-verbose enable log events during process startup/shutdown
-i CONFIG_STRING, inline config which is appended to the config file on-the-fly
--inline-config
--emit-error-log-interval SECONDS
suppress interval seconds of emit error logs
--suppress-repeated-stacktrace [VALUE]
suppress repeated stacktrace
--without-source invoke a fluentd without input plugins
--use-v1-config Use v1 configuration format (default)
--use-v0-config Use v0 configuration format
--strict-config-value Parse config values strictly
--enable-input-metrics Enable input plugin metrics on fluentd
--enable-size-metrics Enable plugin record size metrics on fluentd
-v, --verbose increase verbose level (-v: debug, -vv: trace)
-q, --quiet decrease verbose level (-q: warn, -qq: error)
--suppress-config-dump suppress config dumping when fluentd starts
-g, --gemfile GEMFILE Gemfile path
-G, --gem-path GEM_INSTALL_PATH Gemfile install path (default: $(dirname $gemfile)/vendor/bundle)
--conf-encoding ENCODING specify configuration file encoding
--disable-shared-socket Don't open shared socket for multiple workers
Important Options
-g, --gemfile: Fluentd starts with bundler-managed dependent plugins.
--suppress-config-dump: Fluentd starts without configuration dump. If you do not want to show the configuration in fluentd logs, e.g. it contains private keys, then this option is useful.
--suppress-repeated-stacktrace: If true, suppresses the stacktrace in fluentd logs. Since v0.12, this option is true by default.
--without-source: Fluentd starts without input plugins. This option is useful for flushing buffers with no new incoming events.
-i, --inline-config: If fluentd is used on XaaS which does not support persistent disks, this option is useful.
--no-supervisor: If you want to use your supervisor tools, this option avoids double supervisor.
Set via Configuration File
Some options can be set via <system> directive via configuration file. See configuration file article for more on <system> directive.
fluent-cat
The fluent-cat command sends an event to fluentd in_forward/in_unix plugin. This is particularly useful for testing.
Here is its usage (fluent-cat --help):
Usage: fluent-cat [options] <tag>
-p, --port PORT fluent tcp port (default: 24224)
-h, --host HOST fluent host (default: 127.0.0.1)
-u, --unix use unix socket instead of tcp
-s, --socket PATH unix socket path (default: /var/run/fluent/fluent.sock)
-f, --format FORMAT input format (default: json)
--json same as: -f json
--msgpack same as: -f msgpack
--none same as: -f none
--message-key KEY key field for none format (default: message)
--time-as-integer Send time as integer for v0.12 or earlier
--retry-limit N Specify the number of retry limit (default: 5)
--event-time TIME_STRING Specify the time expression string (default: nil)
Example
Send JSON message with debug.log tag to the local fluentd instance:
echo '{"message":"hello"}' | fluent-cat debug.log
Send JSON message to an instance of fluentd on another machine on the network:
It generates the Fluentd plugin project template. It is good for starting to Fluentd plugin development for using the new API plugin. For more details, refer to the Generating plugin project skeleton section.
Here is its usage (fluent-plugin-generate -h);
Usage: fluent-plugin-generate [options] <type> <name>
Generate a project skeleton for creating a Fluentd plugin
Arguments:
type: input,output,filter,parser,formatter,storage
name: Your plugin name (fluent-plugin- prefix will be added to <name>)
Options:
--[no-]license=NAME Specify license name (default: Apache-2.0)
fluent-ctl
Control fluentd process using Signals or Windows Event.
You can specify the process id of the supervisor process.
fluent-ctl shutdown 11111
If you run Fluentd as a Windows service, then you can omit PID_OR_SVCNAME when the service name is the default value fluentdwinsvc.
fluent-ctl dump
About dump
This function is mainly for Windows.
On Windows, this makes all Fluentd processes (including all worker processes) dump their internal status to the system temp directory (C:\\Windows\\Temp). This is the same behavior as sending SIGCONT to all processes on non-Windows.
Since this uses SIGDUMP, you can change the output path by specifying SIGDUMP_PATH environment variable. Note that the path has to be a file path.
$ $env:SIGDUMP_PATH="/sigdump/sigdump.log" # The directory `sidgump` has to exist.
$ fuentd -c ...
$ fluent-ctl dump # At another shell.
Then Fluentd dumps files as following. Each process id is automatically added to the path.
... [info]: fluent/log.rb:330:info: dump to /sigdump/sigdump-41544.log.
... [info]: #0 fluent/log.rb:330:info: dump to /sigdump/sigdump-21152.log.
... [info]: #1 fluent/log.rb:330:info: dump to /sigdump/sigdump-15656.log.
...
As for non-Windows, you don't have to use this function because you can manually send SIGCONT to each process. Although you can use this function on non-Windows, this just sends SIGCONT to the supervisor process, so you can get only the status of the supervisor process.
fluent-cap-ctl
Control Linux Capability for Fluentd. See Linux Capability article.