Logs

Each Tarantool instance logs important events to its own log file <instance-name>.log. For instances started with tt, the log location is defined by the log_dir parameter in the tt configuration. By default, it’s /var/log/tarantool in the tt system mode, and the var/log subdirectory of the tt working directory in the local mode. In the specified location, tt creates separate directories for each instance’s logs.

To check how logging works, write something to the log using the log module:

$ tt connect my_app
   • Connecting to the instance...
   • Connected to /var/run/tarantool/my_app.control

/var/run/tarantool/my_app.control> require('log').info("Hello for the manual readers")

Then check the logs:

$ tail /var/log/tarantool/my_app.log
2023-09-12 18:13:00.396 [67173] main/111/guard of feedback_daemon/box.feedback_daemon V> metrics_collector restarted
2023-09-12 18:13:00.396 [67173] main/103/-/box.feedback_daemon V> feedback_daemon started
2023-09-12 18:13:00.396 [67173] main/103/- D> memtx_tuple_new_raw_impl(14) = 0x1090077b4
2023-09-12 18:13:00.396 [67173] main/103/- D> memtx_tuple_new_raw_impl(26) = 0x1090077ec
2023-09-12 18:13:00.396 [67173] main/103/- D> memtx_tuple_new_raw_impl(39) = 0x109007824
2023-09-12 18:13:00.396 [67173] main/103/- D> memtx_tuple_new_raw_impl(24) = 0x10900785c
2023-09-12 18:13:00.396 [67173] main/103/- D> memtx_tuple_new_raw_impl(39) = 0x109007894
2023-09-12 18:13:00.396 [67173] main/106/checkpoint_daemon I> scheduled next checkpoint for Tue Sep 12 19:44:34 2023
2023-09-12 18:13:00.396 [67173] main I> entering the event loop
2023-09-12 18:13:11.656 [67173] main/114/console/unix/:/tarantool I> Hello for the manual readers

Log rotation

When logging to a file, the system administrator must ensure logs are rotated timely and do not take up all the available disk space. The recommended way to prevent log files from growing infinitely is using an external log rotation program, for example, logrotate, which is pre-installed on most mainstream Linux distributions.

A Tarantool log rotation configuration for logrotate can look like this:

# /var/log/tarantool/<env>/<app>/<instance>/*.log
/var/log/tarantool/*/*/*/*.log {
    daily
    size 512k
    missingok
    rotate 10
    compress
    delaycompress
    sharedscripts # Run tt logrotate only once after all logs are rotated.
    postrotate
        /usr/bin/tt -S logrotate
    endscript
}

In this configuration, tt logrotate is called after each log rotation to reopen the instance log files after they are moved by the logrotate program.

There is also the built-in function log.rotate(), which you can call on an instance to reopen its log file after rotation.

To learn about log rotation in the deprecated tarantoolctl utility, check its documentation.

Log formats

Tarantool can write its logs to a log file, to syslog, or to a specified program through a pipe.

File is the default log format for tt. To send logs to a pipe or syslog, specify the box.cfg.log parameter, for example:

box.cfg{log = '| cronolog tarantool.log'}
-- or
box.cfg{log = 'syslog:identity=tarantool,facility=user'}

In such configurations, log rotation is usually handled by the external program used for logging.

Version:

Logs

Log rotation

Log formats