pypuppetdb icon indicating copy to clipboard operation
pypuppetdb copied to clipboard

Published 20 hours ago • verified publisher icon voxpupuli

pypuppetdb/types.py: Added noop attribute to Node and Report.

Open jakemalley opened this issue 7 years ago • 5 comments

Added an attribute noop to the Node and Report class types, so it is easy to ascertain whether the node was in noop for a particular report in the case of the Report type, or the last report run in the case of the Node type.

jakemalley avatar Sep 20 '18 16:09 jakemalley

Coverage Status

Coverage increased (+0.04%) to 85.25% when pulling 29fdc64dbb661b71cc0ec506896c3c64c0ac6623 on jakemalley:master into cedeecf48014b4ad5b8e2513ca8230c814f45603 on voxpupuli:master.

coveralls avatar Sep 20 '18 16:09 coveralls

Coverage Status

Coverage increased (+0.04%) to 85.25% when pulling 29fdc64dbb661b71cc0ec506896c3c64c0ac6623 on jakemalley:master into cedeecf48014b4ad5b8e2513ca8230c814f45603 on voxpupuli:master.

coveralls avatar Sep 20 '18 16:09 coveralls

Hi @jakemalley ! Why do we need extra noop fields if in both of these types we already have status with the same information?

gdubicki avatar Mar 21 '21 14:03 gdubicki

Hello @gdubicki, it has been a long time since I've looked at this, but if memory serves it was due to status being set via: 'noop' if noop and noop_pending else status and we wanted to report on whether a system was in 'noop' regardless of the noop_pending flag in certain situations. e.g. we needed to be able to query if the Puppet agent was in 'noop' even if there weren't any changes being applied to the box.

I think also in certain situations if there was a syntax error during catalogue compilation the status was failed (instead of noop) which was triggering false alerts which we wanted to ignore.

jakemalley avatar Mar 21 '21 14:03 jakemalley

I had a feeling that this is because of the way we assign that status now. I am pretty new to the project so I have doubts about the approach we took too. Intuitively for me a run is a noop regardless if it resulted in pending changes or not.

Checking git blame for that 'noop' if noop and noop_pending else status line suggests that we should read #105 -> #103 to learn more about the rationale.

Full disclosure: I haven't done that yet. 😅

...but there seems to be a lot of discussion about a change introduces in PuppetDB 4.1 and being compatible with that. Puppet 4 EOLed in October 2018, so maybe we can simplify the code if we drop 4.x support completely. 🤔

gdubicki avatar Mar 22 '21 17:03 gdubicki