Troubleshooting, Logging, and Support Diagnostics
Logs and levels
The main logs are cogserver.log for server activity, xqe.log for Dynamic Query Mode queries, audit.log for audit events, and bootstrap.log for startup. Log levels range from TRACE and DEBUG through INFO, WARN, ERROR, and FATAL. Set levels in the administration console or configuration, and return them to normal after diagnosis.
Core rules
- Collect diagnostic information from the support/diagnostics feature before restarting; a restart can destroy evidence.
- Use the Dynamic Query Analyzer for XQE query plans and failures.
- Check content store connectivity, schema version, and database credentials when startup fails.
- Audit reporting requires the audit database; without it, only log files remain.
- Correlate timestamps across logs and confirm time synchronization across nodes.
Worked example
Services fail to start after a database migration. The administrator reads bootstrap.log, finds a content store connection error, verifies the JDBC URL and credentials, confirms the schema was upgraded, and restarts. The log level was already INFO, so no change was needed.
Traps
Leaving TRACE enabled in production; restarting before collecting logs; ignoring xqe.log for query errors; and troubleshooting only the application while the database or network is the cause.