How to Modify Warewulf Logging
Introduction
warewulfd can generate massive log output when runtime overlays are transferred—especially across large clusters. This guide shows how to limit that noise by adjusting the log level, helping you preserve disk space and focus on the logs that matter.
Problem
By default, warewulfd logs each request and overlay file transfer at INFO and SERV levels. In clusters with many nodes, this results in gigabytes of log data being generated, making it hard to isolate real issues and consuming disk space rapidly.
Symptoms
The system logs show repetitive entries such as:
wwctl[3151507]: request from hwaddr:xx:xx:xx:xx:xx:xx ipaddr:10.x.x.x | stage:runtime
wwctl[3151507]: SERV : stage_file '/var/lib/warewulf/provision/overlays/node/__RUNTIME__.img'
wwctl[3151507]: send /var/lib/warewulf/provision/overlays/node/__RUNTIME__.img.gz -> node
These log messages appear once per provisioning request, per node.
Resolution
To reduce the logging output, increase the warewulfd log level to WARN (this equates to a number of 30 in the configuration). This filters out routine INFO and SERV messages while still logging warnings and errors.
Step-by-step:
- Set the new log level:
echo WAREWULFD_LOGLEVEL=30 >> /etc/default/warewulfd
- Restart the service:
systemctl restart warewulfd.service
This will apply the new log level setting via the EnvironmentFile used by the systemd service.
Root Cause
The default verbosity includes operational messages that are safe to suppress in most environments. At scale, these messages generate high log volumes unnecessarily.
Notes
The following values are confirmed log level values for warewulf:
- INFO = 20
- SERV = 25
- WARN = 30
Please note - Restarting the service is required for the changes to take effect.