We have a Puppet environment with a few hundred machines. Recently we started seeing agents that stopped working, which results in monitoring alerts. The agents spit out the error message:
Failed to apply catalog: no parameter named 'ensure'
That is the only notification. It does not output a filename in which the parameter is mentioned. Also, there isn't really anything wrong with the manifest. Restarting the puppet agent on the nodes resolves the issue. Sometimes it goes fine for weeks, sometimes it happens twice a day.
The report YAML file contains:
--- !ruby/object:Puppet::Transaction::Report
configuration_version:
logs:
- !ruby/object:Puppet::Util::Log
message: 'Failed to apply catalog: no parameter named ''ensure'''
source: Puppet
tags: !ruby/object:Puppet::Util::TagSet
hash:
err: true
time: 2016-06-02 07:04:56.572138275 +02:00
level: :err
host: X
kind: apply
time: 2016-06-02 07:04:50.366381982 +02:00
status: failed
metrics: {}
transaction_uuid: e9738442-864f-4ebe-81c1-775ef6e6ea35
puppet_version: 4.5.0
environment: production
report_format: 5
resource_statuses: {}
I have no clue what might cause this and I couldn't find anyone with the same problem. Any ideas?
↧