fix: erroneous "new" minion presence events on process startup #64508
Conversation
|
Just as background info, I noticed this because I use the slack engine and presence reactor. |
|
so, this should not be labeled a feature, or it should be going into master. if this is a fix for a bug it should be labeled as such. if it is a new feature it needs to go into master. Next, it is going to need a set of tests and changelog. as well as fixing the pre-commit which just looks like an isort issue. |
|
| so, this should not be labeled a feature, or it should be going into master Good point. I started writing this as a feature but it does appear to be more of a bug fix at this point. I'll swap the title thank you. I'll add tests etc once I get approval on the concept. I don't want to spend time on that if this gets denied |
|
Are these conflicts able to be resolved? |
What does this PR do?
This fixes the issue of the master processes sending incorrect presence events on either the restart of the salt-master process or any of the
Maintenanceprocess restart events (default is every hour). This issue began in 3006.0.Although this was intended to fix the hourly incorrect presence update this also fixes the spammy presence updates between salt-master process restarts.
What issues does this PR fix or reference?
Fixes: #64505
Previous Behavior
The salt-master process sends incorrect presence events at every
restart_intervalof theMaintenanceprocess, signalling that every node connected to the master is "new".New Behavior
Only actual presence events are sent.
Merge requirements satisfied?
[NOTICE] Bug fixes or features added to Salt require tests.
Maintenanceprocess in the current documentation so I think this does not need a specific mention since we are fixing a bugI will update all of these once this is commented on and direction is approved.
Commits signed with GPG?
Yes
Please review Salt's Contributing Guide for best practices.
See GitHub's page on GPG signing for more information about signing commits with GPG.