fix: update log rotation location and frequency (#62)

* fix: update log rotation location

* fix: set up logrotation to execute every hour

This helps avoid situation where postgres logs would fill up the disk
under heavy usage conditions. Technically, it can still happen, but
should be exceedingly rare outside of abusive conditions.

Author: darora

ansible/files/logrotate-postgres

@@ -1,10 +1,9 @@
-/var/lib/postgresql/12/main/pg_log/postgresql.csv {
-    daily
-    size 50M 
+/var/lib/postgresql/data/pg_log/postgresql.log {
+    size 50M
     rotate 3  
     copytruncate
     delaycompress
     compress
     notifempty
     missingok
-}
+}

ansible/playbook.yml

@@ -96,6 +96,17 @@
         src: files/logrotate-postgres
         dest: /etc/logrotate.d/postgres
 
+    - name: Configure logrotation to run every hour
+      shell:
+        cmd: mv /etc/cron.daily/logrotate /etc/cron.hourly/
+      become: yes
+
+    - name: restart crond
+      systemd:
+        state: restarted
+        name: cron
+      become: yes
+
     - name: Enhance fail2ban
       import_tasks: tasks/setup-fail2ban.yml