Skip to main content
Question

YAML Files Flooding Var Directory Every Few Minutes

  • November 10, 2025
  • 1 reply
  • 22 views
P

Hello evb,

in a 3-host Nutanix Cluster (AOS 6.10,1 and AHV v.20230302.103003) I have one of the 3 hosts with the [/var] filled up at 92% (2.5GB on 2.9GB)

The cluster is up’n’running and was OK until this warning raised by the ncc checks (v 5.1.0)

Most of the space is used by:

6565632 /var/log
4502260 /var/log/sa
2546588 /var/lib
2406748 /var/lib/puppet
2406072 /var/lib/puppet/reports
2241772 /var/lib/puppet/reports/HOSTNAME

In the last dir I have more than 90K [.YAML] files (of about 20KB each and created every 2 or 3 minutes, starting from the first host setup), while in the other 2 hosts I have only 14 [.yaml] files.

File names are in the form 

yyyymmddhhmm.yaml

What are they used for?

Is it safe to delete the oldest ones?

TIA

Paolo

 

1 reply

aluciani
Forum|alt.badge.img+34
  • aluciani
  • Chevalier
  • 338 replies
  • November 14, 2025

Hey ​@PAP my thoughts

Try this in your dev or staging env before trying on production.

Why is this happening?

  • Puppet typically runs every few minutes to enforce configuration state.
  • Each run generates a report file under /var/lib/puppet/reports/.
  • If reporting cleanup is not configured, these files will keep growing.

What can you do?

  • Verify Puppet Run Interval:
  • Check the Puppet agent configuration (/etc/puppet/puppet.conf) for the runinterval setting. Default is 30 minutes, but if it’s set lower, reports will accumulate faster.

Enable Report Cleanup:

  • Puppet does not automatically purge old reports. You can:
  • Use the puppet reportcleaner module or a cron job to remove old files.

Check NCC Warning:

  • The NCC alert is triggered because /var is filling up. Cleaning old reports is safe as they are only historical logs of Puppet runs.


Terms & ConditionsAccessibility statement