timers continue when freezing operations

[rtoma]

Long story short

Thanks for creating kopf. I am currently developing an operator that updates Egress NetworkPolicy resources to sync its spec.egress[*].to[*].ipBlock.cidr keys with A records for DNS hostnames.

This operator used timers (looking into daemons to follow DNS TTL). I run an operator in k8s with cluster peering. For development I start a local instance with --dev.

I see the operator in k8s detects the development instance:

Freezing operations in favour of [<Peer rtoma@devsystem/20210208145159/nz5: priority=666, lifetime=60, lastseen='2021-02-08T14:52:01.160123'>].

The development instance's log show:

Resuming operations after the freeze. Conflicting operators with the same priority are gone.

Still both operator instances show Timer 'NAME' succeeded messages. I searched the docs, but could not find anything.

Description

The code snippet to reproduce the issue

import kopf

The exact command to reproduce the issue

kopf run ...

The full output of the command that failed

Environment

• Kopf version: 1.29.1
• Kubernetes version: 1.19.5
• Python version: 3.9.1
• OS/platform: k8s on pi4, local on arm mbp

Python packages installed

[nolar]

Indeed, that’s possible. And that is clearly a bug. Thanks for reporting — I’ll take a look if it can be fixed easily.

[nolar]

The fix is released in kopf>=1.30.0rc1.