Troubleshooting: Polling Not Working

Symptom

The system shows a warning message:

No poll activity has been recorded yet; verify the poller service is running and configured.

This means the poller service hasn't successfully written any poll records to the database.

Quick Diagnostic Steps

Run these commands to quickly identify the issue:

# 1. Check if poller service is running
sudo systemctl status eas-station-poller.service

# 2. Check recent logs
sudo journalctl -u eas-station-poller.service -n 100 --no-pager

# 3. Check if service is enabled
sudo systemctl is-enabled eas-station-poller.service

# 4. Verify database connection
sudo -u easstation psql -h localhost -U eas_station -d alerts -c "SELECT COUNT(*) FROM poll_history;"

Common Issues and Solutions

Issue 1: Service Not Running

Symptoms:

$ sudo systemctl status eas-station-poller.service
● eas-station-poller.service - EAS Station™ Alert Poller
   Loaded: loaded (/etc/systemd/system/eas-station-poller.service; enabled)
   Active: inactive (dead)

Solution:

# Start the service
sudo systemctl start eas-station-poller.service

# Enable it to start on boot
sudo systemctl enable eas-station-poller.service

# Check status again
sudo systemctl status eas-station-poller.service

Issue 2: Service Failing to Start

Symptoms:

$ sudo systemctl status eas-station-poller.service
● eas-station-poller.service - EAS Station™ Alert Poller
   Loaded: loaded (/etc/systemd/system/eas-station-poller.service; enabled)
   Active: failed (Result: exit-code)

Diagnosis:

# Check recent logs for errors
sudo journalctl -u eas-station-poller.service -n 200 --no-pager

# Look for these common errors:
# - "could not connect to server" → Database connection issue
# - "Permission denied" → File/directory permissions
# - "No module named" → Python dependency missing
# - "No endpoints configured" → Configuration missing

Common Fixes:

A. Database Connection Failed

# Error: "could not connect to server: Connection refused"
# Cause: PostgreSQL not running or wrong credentials

# Check PostgreSQL is running
sudo systemctl status postgresql

# Start PostgreSQL if needed
sudo systemctl start postgresql

# Verify database credentials in .env
sudo nano /opt/eas-station/.env
# Check: POSTGRES_HOST, POSTGRES_PORT, POSTGRES_USER, POSTGRES_PASSWORD

# Test connection manually
sudo -u easstation psql -h localhost -U eas_station -d alerts -c "\dt"

B. Python Module Missing

# Error: "No module named 'requests'" (or other module)
# Cause: Python dependencies not installed

# Reinstall requirements
sudo -u easstation pip3 install -r /opt/eas-station/requirements.txt

# Restart service
sudo systemctl restart eas-station-poller.service

C. No Endpoints Configured

Error: "No endpoints configured"
Cause: Missing IPAWS feed URLs and/or zone codes

Polling endpoints, the NOAA user-agent string, and location/zone codes are now configured through the admin UI, not .env:

• Poll interval, CAP timeout, NOAA user-agent, IPAWS feed URLs, lookback hours — Settings → Poller (/admin/poller, persisted in the poller_settings table).
• County, state, FIPS codes, NWS zone codes — Settings → Location and Settings → Alert Filtering (/admin/location_settings and /admin/alert_filtering).

After saving, restart the poller so it picks up the new values:

sudo systemctl restart eas-station-poller.service

Issue 3: Database Permission Denied

Symptoms:

ERROR: permission denied for table poll_history

Solution:

# Grant permissions to eas_station user
sudo -u postgres psql -d alerts << EOF
GRANT ALL PRIVILEGES ON ALL TABLES IN SCHEMA public TO eas_station;
GRANT ALL PRIVILEGES ON ALL SEQUENCES IN SCHEMA public TO eas_station;
GRANT USAGE, SELECT ON ALL SEQUENCES IN SCHEMA public TO eas_station;
EOF

# Restart poller
sudo systemctl restart eas-station-poller.service

Issue 4: Service Running But Not Polling

Symptoms:

• Service status shows "active (running)"
• But no poll records appear in database
• No errors in logs

Diagnosis:

# Check if poller is actually making requests
sudo journalctl -u eas-station-poller.service -f

# You should see lines like:
# "Polling: https://apps.fema.gov/IPAWSOPEN_EAS_SERVICE/rest/public/recent/..."
# "NOAA endpoint: https://api.weather.gov/alerts/active?zone=..."

# If you see nothing, the poller might be stuck or waiting

Solutions:

A. Check the poll interval

Default is 120 seconds (2 minutes).
If set too high, you'll wait a long time to see activity.

The poll interval is now stored in the database (poller_settings.poll_interval_sec) and configured at Settings → Poller (/admin/poller). To inspect it from the command line:

sudo -u easstation psql -h localhost -U eas_station -d alerts \
  -c "SELECT poll_interval_sec FROM poller_settings;"
# Should show 120 (or your configured value)

If the value is something like 3600 (1 hour), that's why you're not seeing polls — lower it via the Poller settings page.

B. Force Manual Poll

# Run poller once manually to test
sudo -u easstation /usr/bin/python3 /opt/eas-station/poller/cap_poller.py \
  --database-url "postgresql://eas_station:YOUR_PASSWORD@localhost:5432/alerts" \
  --once

# Check for errors in output

C. Check Network Connectivity

# Test IPAWS feed access
curl -I "https://apps.fema.gov/IPAWSOPEN_EAS_SERVICE/rest/public/recent/2025-01-01T00:00:00-00:00"

# Should return: HTTP/2 200

# Test NOAA API access
curl -I "https://api.weather.gov/alerts/active?zone=OHZ016"

# Should return: HTTP/2 200

# If these fail, check firewall/network

Issue 5: Wrong Database Credentials in Service File

Symptoms:

FATAL: password authentication failed for user "postgres"

Root Cause: The systemd service file uses environment variables from .env, but if POSTGRES_USER is set differently in .env vs what the service expects, authentication fails.

Diagnosis:

# Check what the service file expects
grep "ExecStart" /etc/systemd/system/eas-station-poller.service

# Should show:
# ExecStart=/usr/bin/python3 /opt/eas-station/poller/cap_poller.py --continuous --database-url postgresql://postgres:${POSTGRES_PASSWORD}@${POSTGRES_HOST}:${POSTGRES_PORT}/${POSTGRES_DB}

# Note: It hardcodes "postgres" as the user!
# But your .env might have POSTGRES_USER=eas_station

Solution:

Option A: Update service file to use correct user

# Edit service file
sudo nano /etc/systemd/system/eas-station-poller.service

# Change line 18 from:
ExecStart=/usr/bin/python3 /opt/eas-station/poller/cap_poller.py --continuous --database-url postgresql://postgres:${POSTGRES_PASSWORD}@${POSTGRES_HOST}:${POSTGRES_PORT}/${POSTGRES_DB}

# To:
ExecStart=/usr/bin/python3 /opt/eas-station/poller/cap_poller.py --continuous --database-url postgresql://${POSTGRES_USER}:${POSTGRES_PASSWORD}@${POSTGRES_HOST}:${POSTGRES_PORT}/${POSTGRES_DB}

# Reload systemd
sudo systemctl daemon-reload

# Restart service
sudo systemctl restart eas-station-poller.service

Option B: Change .env to match service file

# Edit .env
sudo nano /opt/eas-station/.env

# Change:
POSTGRES_USER=eas_station

# To:
POSTGRES_USER=postgres

# And grant permissions to postgres user
sudo -u postgres psql -d alerts << EOF
GRANT ALL PRIVILEGES ON ALL TABLES IN SCHEMA public TO postgres;
GRANT ALL PRIVILEGES ON ALL SEQUENCES IN SCHEMA public TO postgres;
EOF

# Restart service
sudo systemctl restart eas-station-poller.service

Verification

After fixing the issue, verify polling is working:

# 1. Check service is running and healthy
sudo systemctl status eas-station-poller.service

# 2. Watch logs in real-time (wait at least one poll interval; default 120s)
sudo journalctl -u eas-station-poller.service -f

# You should see:
# "[CAP_POLLER] Starting continuous polling..."
# "Polling: https://apps.fema.gov/..."
# "Processed 0-X alerts from IPAWS"
# "Poll completed successfully"

# 3. Check database for poll records
sudo -u easstation psql -h localhost -U eas_station -d alerts -c "SELECT timestamp, status, data_source, alerts_fetched FROM poll_history ORDER BY timestamp DESC LIMIT 5;"

# Should show recent poll records with timestamps

Prevention

Enable Service on Boot

# Ensure poller starts automatically
sudo systemctl enable eas-station-poller.service

# Verify
sudo systemctl is-enabled eas-station-poller.service
# Should show: enabled

Monitor Poller Health

Add to cron to check poller health:

# Edit crontab
crontab -e

# Add (checks every 15 minutes):
*/15 * * * * systemctl is-active --quiet eas-station-poller.service || systemctl start eas-station-poller.service

Check Logs Regularly

# Add alias to .bashrc for easy log checking
echo 'alias poller-logs="sudo journalctl -u eas-station-poller.service -n 100 --no-pager"' >> ~/.bashrc
source ~/.bashrc

# Now you can just run:
poller-logs

Related Documentation

• IPAWS Feed Integration Guide
• Installation Details

Still Not Working?

If none of these solutions work, gather diagnostic information:

# Create diagnostics bundle
mkdir -p /tmp/eas-diagnostics

# Service status
sudo systemctl status eas-station-poller.service > /tmp/eas-diagnostics/service-status.txt

# Recent logs
sudo journalctl -u eas-station-poller.service -n 500 --no-pager > /tmp/eas-diagnostics/poller-logs.txt

# Configuration (with passwords redacted)
grep -v "PASSWORD\|SECRET\|KEY" /opt/eas-station/.env > /tmp/eas-diagnostics/config.txt

# Database check
sudo -u easstation psql -h localhost -U eas_station -d alerts -c "\dt" > /tmp/eas-diagnostics/db-tables.txt 2>&1

# Manual test run
sudo -u easstation /usr/bin/python3 /opt/eas-station/poller/cap_poller.py --once --database-url "postgresql://eas_station:REDACTED@localhost:5432/alerts" > /tmp/eas-diagnostics/manual-run.txt 2>&1

# Package everything
tar -czf /tmp/eas-diagnostics.tar.gz -C /tmp eas-diagnostics/

echo "Diagnostics saved to /tmp/eas-diagnostics.tar.gz"
echo "Share this file when requesting support"