When setting up runit logging for a service, you might encounter situations where your log directory remains empty despite proper configuration. The service appears to run normally according to sv status, but the expected log files never appear in your designated directory.

A standard runit service with logging consists of these components:

/etc/sv/test/
├── run
├── finish
└── log/
    ├── run
    └── config

The most common issue lies in the log/run script configuration. Many developers make these mistakes:

1. Forgetting the exec command
2. Not properly piping service output
3. Incorrect permission settings

Here's the corrected version that actually works:

#!/bin/sh
exec svlogd -tt /var/log/test

Let me show a fully functional setup that handles both service management and logging:

Main service run script

#!/bin/sh
exec 2>&1
exec /path/to/your/service/binary 2>&1 | logger -t yourservice

Log directory run script

#!/bin/sh
exec chpst -ulogger svlogd -tt /var/log/test

When logs don't appear, follow this checklist:

1. Verify all scripts are executable: chmod +x /etc/sv/test/*/run
2. Check ownership of log directory: chown logger:logger /var/log/test
3. Test log script manually: /etc/sv/test/log/run
4. Examine process tree: ps auxf | grep -A5 svlogd

For better log management, create a config file in your log directory:

# /etc/sv/test/log/config
s1000000
n10
t86400
!gzip

This configuration will:

• Rotate logs at 1MB (s1000000)
• Keep 10 archived logs (n10)
• Compress old logs (!gzip)
• Add timestamps (t86400 for daily rotation)

Remember these essential commands for managing runit services with logging:

# Check status
sv status test
sv status test/log

# Start/stop logging separately
sv up test/log
sv down test/log

# Full service control
sv restart test
sv force-stop test

---

Runit's logging system works through a separate logging service that runs parallel to your main service. The key components are:

/etc/sv/your_service/
├── run
├── log/
│   ├── run
│   └── config

From your description, several potential issues emerge:

1. The logging service isn't properly starting (shown as "down: log")
2. Potential permission issues with the log directory
3. Configuration problems in the log/run script

Here's a fully working configuration example:

# /etc/sv/test/run
#!/bin/sh
exec 2>&1
exec your_service_command

# /etc/sv/test/log/run
#!/bin/sh
exec chpst -u logger svlogd -tt /var/log/test

1. Permission setup:

sudo mkdir -p /var/log/test
sudo chown -R logger:logger /var/log/test
sudo chmod 755 /var/log/test

2. Log rotation configuration (optional):

# /etc/sv/test/log/config
s1000000
n10
t86400
!gzip

Check service status:

sv status test
sv status test/log

Force restart both services:

sv restart test
sv restart test/log

For applications that need special log handling:

# Alternative log/run script using multilog
#!/bin/sh
exec chpst -u logger multilog t ./main

For syslog integration:

#!/bin/sh
exec chpst -u logger svlogd -tt /var/log/test | logger -t test-service