Where Are PostgreSQL Logs Stored? Default Paths (Ubuntu, CentOS, Docker)

PostgreSQL's log location is not fixed — it depends on three settings: log_destination, logging_collector, and log_directory. On Debian and Ubuntu the package enables file logging at /var/log/postgresql/postgresql-<version>-<cluster>.log. On RHEL and CentOS the logging collector writes to the log/ subdirectory inside the data directory, typically /var/lib/pgsql/<version>/data/log/. When logging_collector is off and log_destination is stderr (the upstream default and the Docker default), output goes to the process's stderr — captured by the systemd journal or by docker logs.

Quick Reference: PostgreSQL Log File Paths

How to View and Tail PostgreSQL Logs

On Debian/Ubuntu, find and tail the current file:

sudo tail -f /var/log/postgresql/postgresql-*.log

When using the systemd journal:

sudo journalctl -u postgresql -f
sudo journalctl -u postgresql@16-main -f   # versioned unit

Ask PostgreSQL for the exact active file:

SELECT pg_current_logfile();

Grep for errors and fatals:

sudo grep -E 'ERROR|FATAL|PANIC' /var/log/postgresql/postgresql-*.log

In Docker:

docker logs -f <container>

Finding the Exact Path When It Varies

The authoritative answer comes from the running server:

SHOW log_destination;     -- stderr, csvlog, jsonlog, syslog (or a combination)
SHOW logging_collector;   -- on means PostgreSQL manages log files itself
SHOW data_directory;      -- e.g. /var/lib/pgsql/16/data
SHOW log_directory;       -- absolute, or relative to data_directory
SHOW log_filename;        -- e.g. postgresql-%Y-%m-%d_%H%M%S.log

If logging_collector is on, the file lives at log_directory (joined to data_directory when relative). If it's off and log_destination is stderr, look in the journal or docker logs.

Ubuntu / Debian Detail

Debian uses a cluster system. Config lives at /etc/postgresql/<version>/<cluster>/postgresql.conf, the data directory at /var/lib/postgresql/<version>/<cluster>, and logs at /var/log/postgresql/. The package ships with logging_collector effectively redirecting to those files. Edit settings, then reload or restart:

sudo -u postgres psql -c "SELECT pg_reload_conf();"
sudo systemctl restart postgresql@16-main   # for changes requiring restart

RHEL / CentOS / Fedora Detail

Config and data both live under /var/lib/pgsql/<version>/data/, with logs in the log/ subdirectory. The unit is postgresql-<version>. SELinux confines the postgres process, so relocating logs outside the data directory needs a matching file context (postgresql_log_t).

Docker Detail

The official postgres image logs to stderr, so docker logs <container> is the default way to read them and there is no file inside the container. To write files, mount a volume and pass logging flags:

docker run -d --name pg \
  -v /host/pglog:/var/log/postgresql \
  -e POSTGRES_PASSWORD=secret postgres \
  -c logging_collector=on \
  -c log_directory=/var/log/postgresql \
  -c log_filename=postgresql.log

Changing the Log Location and Verbosity

Key parameters in postgresql.conf:

log_destination = 'stderr'          # or 'csvlog', 'jsonlog', combine with commas
logging_collector = on              # required for PostgreSQL-managed files (needs restart)
log_directory = 'log'               # relative to data_directory, or absolute
log_filename = 'postgresql-%Y-%m-%d_%H%M%S.log'
log_min_duration_statement = 250    # log queries slower than 250 ms
log_line_prefix = '%m [%p] %q%u@%d ' # timestamp, pid, user@db
log_statement = 'ddl'               # none, ddl, mod, or all

logging_collector requires a full restart. Most other logging parameters take effect with SELECT pg_reload_conf();.

Log Rotation

PostgreSQL's built-in collector rotates on its own when log_rotation_age (default 1d) or log_rotation_size (default 10MB) is reached, naming files via log_filename patterns. With log_truncate_on_rotation = on plus a time-based log_filename (e.g. postgresql-%a.log for day-of-week), it cycles through a fixed weekly set. On Debian, an external rotation also exists at /etc/logrotate.d/postgresql-common:

sudo logrotate -f /etc/logrotate.d/postgresql-common

Use either the built-in collector rotation or logrotate — not both on the same files.

Common Errors You'll Find in PostgreSQL Logs

• FATAL: password authentication failed for user "x" — wrong password or a pg_hba.conf rule using md5/scram the client can't satisfy.
• FATAL: no pg_hba.conf entry for host "1.2.3.4" — the client IP/database/user combo isn't allowed; add a rule and reload.
• ERROR: relation "foo" does not exist — wrong schema/search_path or the table truly is missing.
• FATAL: sorry, too many clients already — connections hit max_connections; add pooling (PgBouncer) or raise the limit.
• could not extend file ... No space left on device — the data volume is full; free space or grow the disk before the DB stops accepting writes.
• deadlock detected — two transactions blocked each other; PostgreSQL aborted one. Review the logged statements to fix lock ordering.

Troubleshooting: Logs Missing or Empty

• Collector off — with logging_collector = off, there is no file; check journalctl -u postgresql or docker logs.
• Wrong directory — log_directory may be relative to data_directory; resolve it with SHOW data_directory; + SHOW log_directory;, or just SELECT pg_current_logfile();.
• Verbosity too low — log_min_messages/log_statement may be filtering out what you expect; raise them and pg_reload_conf().
• Permissions — files in the data directory are owned by the postgres user (mode 600); use sudo -u postgres or root.
• Server won't start — startup failures (bad config, port in use) appear in the journal even when file logging is configured, because the collector isn't running yet.
• Docker — no file inside the image unless you enabled the collector and mounted a volume; otherwise use docker logs.