Logs

The system maintains a comprehensive list of logs under /var/log:

  • The platform/ directory has logs pertaining the the general platform or to specific log types.
    • apps.log contains application and process control logging
    • audit.log a log type available if enabled with log audit locallog on - contains audit log information
    • database.log contains database logs spawned by VOSS-4-UC transactions
    • dockerd.log contains logs spawned by the Docker container management daemon
    • backup.log contains all logging pertaining to backups - available after the first backup
    • cluster.log contains all control level management of the cluster
    • config.log contains information relating to the platform-level configuration
    • event.log a log type available if enabled with log event locallog on - contains event log information
    • execute.log contains low-level information about command execution
    • notifications.log contains information relating to SNMP notifications
    • reports.log contains information relating to system reports. Refer to the Scheduling section on how reports can be created.
    • security.log contains information relating to user security
    • security_install_GUI.log contains information relating to user graphical interface security
    • security_update.log contains information relating to security updates
    • ui.log contains higher-level information relating to UI commands being executed.
    • wsgi.log contains information relating to API-level commands via the WSGI server
  • The provision/ directory contains logs relating to provisioning. Every module provision is logged to component log files.
  • The health/ directory contains health logs. These are stored automatically every half hour, or whenever health is run, and are of the format health/summary_report-<date>-<time>.
  • The process/ directory contains process logs instrumental in debugging particular processes. All of the output from each process is logged to an individual file process/<application>.<process>.log
  • The install/ directory contains logs detailing the install process.
  • The mongodb/ directory contains logs relating to the Database function.
  • The nginx/ directory contains logs relating to the WebProxy function.
  • The voss-deviceapi/ directory contains logs relating to the Application function. For example, voss-deviceapi/cnf_collector.log is the Change Notification Collector log.

log list [<search_string>] is used to display a list of logs, optionally matching search_string. For example:

platform@clusternode:~$ log list alternatives.log
selfservice/alternatives.log
voss-deviceapi/alternatives.log
nginx/alternatives.log
mongodb/alternatives.log
alternatives.log

The main log rotation scripts will rotate log files only when files exceed 100M or if the disk containing /var/log/ is over 80% full. This is checked once per hour. The system will attempt to keep 5 historic zipped files of each log. If the disk containing /var/log is over 90% full, files will be purged to ensure that the system continues to function.

100M size logs:

  • mongodb/*.log
  • nginx/*.log
  • selfservice/*.log
  • voss-deviceapi/*.log

10M size logs:

  • other logs in /var/log/ and sub-directories not specified above

All rotated log files and log files exceeding 1GB can be purged manually with log purge.