All of my hosts are managed by puppet but for the most of them, the reports they send are not identified as puppet reports. But few of them are working properly.
When reporting a failed state, the origin is correctly flagged as Puppet .
I have this behaviour on my 3 Foreman servers (test, qualifying and production)
I migrate from CentOS7 to OracleLinux 8 earlier this year, I don’t know if it’s linked or not.
Thanks for helping
Correctly identifing puppet reports
Foreman and Proxy versions:
Foreman and Proxy plugin versions:
Distribution and version:
Other relevant data:
We run into the same problem.
As a workaround you can change the following file (but I dont know whats the impact):
- next if log.message =~ /^(Finished catalog run|Applied catalog) in \d+.\d+ seconds$/
+ next if log.message =~ /^Finished catalog run in \d+.\d+ seconds$/
Thanks for you response @chr1s !
Have this bug been signaled to the Foreman team ?
It would be nice to evaluate the impact and release an “official” fix
The upgrade docs for Foreman 3.5 do mention this issue under “Upgrade Warnings”, there are also links to the related issues in that segment. So yes, this is known and worked on.
The “fix” proposed by @chr1s is just reverting the change that caused this behavior, which was originally inteded to reduce database storage space by filtering out unnecessary lines from the report.
This is tracked in:
I should have reverted earlier, but just did that in:
It’s in Foreman 3.7, which is in the process of being released: Foreman 3.7.0 (GA) release process
Once it’s released then Foreman 3.5 will be end of life, but we can look at doing a patch release for Foreman 3.6.
I missed it!
It seems it’s on a good way to be resolved soon.