Added weekly VACUUM FULL maintenance with cron scheduling for database optimization

openremote/openremote#1423

Author: Miggets7

Dockerfile

@@ -3,10 +3,13 @@ ARG PG_MAJOR_PREVIOUS=14
 ARG PG_MAJOR=15
 
 FROM timescaledev/timescaledb-ha:pg15-multi as trimmed
-MAINTAINER [email protected]
+LABEL maintainer="[email protected]"
 
 USER root
 
+# Install cron for scheduled VACUUM FULL operations
+RUN apt-get update && apt-get install -y cron && rm -rf /var/lib/apt/lists/*
+
 # Give postgres user the same UID and GID as the old alpine postgres image to simplify migration of existing DB
 RUN usermod -u 70 postgres \
  && groupmod -g 70 postgres \

Dockerfile.alpine

@@ -20,7 +20,7 @@ FROM timescaledev/timescaledb-ha:pg15-multi as timescale-ha
 # This still doesn't work as timescale code is compiled against glibc and some references don't match with gcompat
 
 FROM timescale/timescaledb:latest-pg15
-MAINTAINER [email protected]
+LABEL maintainer="[email protected]"
 
 ENV GLIBC_VERSION 2.35-r0
 ENV TZ ${TZ:-Europe/Amsterdam}
@@ -29,8 +29,8 @@ ENV POSTGRES_DB ${POSTGRES_DB:-openremote}
 ENV POSTGRES_USER ${POSTGRES_USER:-postgres}
 ENV POSTGRES_PASSWORD ${POSTGRES_PASSWORD:-postgres}
 
-# Add glibc
-RUN apk add gcompat
+# Add glibc and cron for scheduled VACUUM FULL operations
+RUN apk add --no-cache gcompat dcron
 
 
 COPY --from=timescale-ha /usr/lib/postgresql/15/lib/bitcode/postgis-3/ /usr/local/lib/postgresql/bitcode/

docker-entrypoint-initdb.d/200_or_setup_vacuum_full.sh

@@ -0,0 +1,63 @@
+#!/bin/sh
+
+# This script sets up a weekly VACUUM FULL job to clean up database bloat
+# VACUUM FULL reclaims more space than regular AUTOVACUUM by rewriting the entire table
+
+# Create scripts directory if it doesn't exist
+mkdir -p /var/lib/postgresql/scripts
+
+# Create the vacuum script that will run VACUUM FULL on all tables in the database
+cat > /var/lib/postgresql/scripts/vacuum_full.sh << 'EOF'
+#!/bin/sh
+
+# Log file for vacuum operations
+LOG_FILE="/var/lib/postgresql/vacuum_full.log"
+
+echo "$(date): Starting weekly VACUUM FULL operation" >> "$LOG_FILE"
+
+# Get list of all tables in the database
+TABLES=$(psql -t -c "SELECT schemaname || '.' || tablename FROM pg_tables WHERE schemaname NOT IN ('pg_catalog', 'information_schema')" "${POSTGRES_DB:-postgres}")
+
+# Run VACUUM FULL on each table
+for TABLE in $TABLES; do
+    echo "$(date): Running VACUUM FULL on $TABLE" >> "$LOG_FILE"
+    psql -c "VACUUM FULL $TABLE" "${POSTGRES_DB:-postgres}"
+    echo "$(date): Completed VACUUM FULL on $TABLE" >> "$LOG_FILE"
+done
+
+echo "$(date): Weekly VACUUM FULL operation completed" >> "$LOG_FILE"
+EOF
+
+# Make the script executable
+chmod +x /var/lib/postgresql/scripts/vacuum_full.sh
+
+# Set up cron job to run the script weekly (Sunday at 12:00 AM)
+CRON_ENTRY="0 0 * * 0 /var/lib/postgresql/scripts/vacuum_full.sh"
+
+# Detect which cron system is available
+if command -v crond > /dev/null 2>&1; then
+    # Alpine Linux (dcron)
+    echo "Setting up cron job using Alpine's crond (dcron)..."
+    mkdir -p /etc/crontabs
+    echo "$CRON_ENTRY" > /etc/crontabs/postgres
+    chmod 600 /etc/crontabs/postgres
+    echo "Weekly VACUUM FULL job has been scheduled (Sundays at 12:00 AM) using dcron"
+    
+    # Create a script to start crond when container starts
+    cat > /var/lib/postgresql/scripts/start_cron.sh << 'EOF'
+#!/bin/sh
+if [ -x "$(command -v crond)" ]; then
+  echo "Starting crond service for scheduled database maintenance..."
+  crond -b -L /var/lib/postgresql/cron.log
+fi
+EOF
+    chmod +x /var/lib/postgresql/scripts/start_cron.sh
+    
+elif command -v cron > /dev/null 2>&1; then
+    # Debian/Ubuntu (standard cron)
+    echo "Setting up cron job using standard cron..."
+    (crontab -u postgres -l 2>/dev/null || echo "") | echo "$CRON_ENTRY" | crontab -u postgres -
+    echo "Weekly VACUUM FULL job has been scheduled (Sundays at 12:00 AM) using standard cron"
+else
+    echo "WARNING: No cron system found. VACUUM FULL will not be automatically scheduled."
+fi

or-entrypoint.sh

@@ -377,4 +377,13 @@ if [ -n "$DATABASE_ALREADY_EXISTS" ]; then
     fi
 fi
 
+# Start cron service if it exists (for scheduled VACUUM FULL operations)
+if [ -x "$(command -v cron)" ]; then
+  echo "Starting cron service for scheduled database maintenance..."
+  service cron start
+elif [ -x "$(command -v crond)" ] && [ -f "/var/lib/postgresql/scripts/start_cron.sh" ]; then
+  echo "Starting Alpine crond service for scheduled database maintenance..."
+  /var/lib/postgresql/scripts/start_cron.sh
+fi
+
 exec /docker-entrypoint.sh $@