Handler not being called

[jaepetto]

Keywords

handler

Problem

I have a very simple script that uses Kopf. It is triggered when a Workflow is being created. This handler, in turn, generates a "bmc.tinkerbell.org/v1alpha1/Job" object in Kubernetes (without adopting it). Now, I want to create a second handler in the same file that handles the update of the said Job:

@kopf.on.create("Workflow", when=should_trigger_creation)
def create_fn(body, logger, **_):
    ...
    return object_name

@kopf.on.update("bmc.tinkerbell.org", "v1alpha1", "Job")
def on_bmc_job_update_fn(spec, old, new, diff, logger, **_):
    logger.info(f"---> new: {new}")
    logger.info(f"---> diff: {diff}")

Surprisingly, only the first handler (create_fn) is ever being called. The handler called on_bmc_job_update_fn is actually never called.

When looking at the kopf logs, I can see that only one handler is registered:

/usr/local/lib/python3.12/site-packages/kopf/_core/reactor/running.py:179: FutureWarning: Absence of either namespaces or cluster-wide flag will become an error soon. For now, switching to warnings.warn("Absence of either namespaces or cluster-wide flag will become an error soon." [2024-11-21 13:53:12,606] kopf._core.reactor.r [DEBUG ] Starting Kopf 1.37.2. [2024-11-21 13:53:12,606] kopf._core.engines.a [INFO ] Initial authentication has been initiated. [2024-11-21 13:53:12,606] kopf.activities.auth [DEBUG ] Activity 'login_via_client' is invoked. [2024-11-21 13:53:12,608] kopf.activities.auth [DEBUG ] Client is configured in cluster with service account. [2024-11-21 13:53:12,609] kopf.activities.auth [INFO ] Activity 'login_via_client' succeeded. [2024-11-21 13:53:12,609] kopf._core.engines.a [INFO ] Initial authentication has finished. [2024-11-21 13:53:12,724] kopf._cogs.clients.w [DEBUG ] Starting the watch-stream for customresourcedefinitions.v1.apiextensions.k8s.io cluster-wide. [2024-11-21 13:53:12,725] kopf._cogs.clients.w [DEBUG ] Starting the watch-stream for workflows.v1alpha1.tinkerbell.org cluster-wide. [2024-11-21 13:53:12,725] kopf._cogs.clients.w [DEBUG ] Starting the watch-stream for jobs.v1alpha1.bmc.tinkerbell.org cluster-wide. [2024-11-21 13:53:51,323] kopf.objects [DEBUG ] [tink/icadmin012] Creation is in progress: {'apiVersion': 'tinkerbell.org/v1alpha1', 'kind': 'Workflow', 'metadata': {'annotations' [2024-11-21 13:53:51,323] kopf.objects [DEBUG ] [tink/icadmin012] Handler 'create_fn' is invoked. [2024-11-21 13:53:51,324] kopf.objects [INFO ] [tink/icadmin012] Creating boot-to-PXE object for hardware: icadmin012 [2024-11-21 13:53:51,340] kubernetes.client.re [DEBUG ] response body: {"apiVersion":"bmc.tinkerbell.org/v1alpha1","kind":"Job","metadata":{"creationTimestamp":"2024-11-21T13:53:51Z","gen

[2024-11-21 13:53:51,340] kopf.objects [DEBUG ] [tink/icadmin012] Boot to PXE child is created: {'apiVersion': 'bmc.tinkerbell.org/v1alpha1', 'kind': 'Job', 'metadata': {'creation [2024-11-21 13:53:51,341] kopf.objects [INFO ] [tink/icadmin012] object_name: {'boot-to-pxe-name': 'reboot-icadmin012-to-pxe'} [2024-11-21 13:53:51,343] kopf.objects [INFO ] [tink/icadmin012] Handler 'create_fn' succeeded. [2024-11-21 13:53:51,343] kopf.objects [INFO ] [tink/icadmin012] Creation is processed: 1 succeeded; 0 failed. [2024-11-21 13:53:51,344] kopf.objects [DEBUG ] [tink/icadmin012] Patching with: {'status': {'create_fn': {'boot-to-pxe-name': 'reboot-icadmin012-to-pxe'}}, 'metadata': {'annotati [2024-11-21 13:53:51,368] kopf.objects [WARNING ] [tink/icadmin012] Patching failed with inconsistencies: (('remove', ('status', 'create_fn'), {'boot-to-pxe-name': 'reboot-icadmin01 [2024-11-21 13:53:51,461] kopf.objects [DEBUG ] [tink/reboot-icadmin012-to-pxe] Creation is in progress: {'apiVersion': 'bmc.tinkerbell.org/v1alpha1', 'kind': 'Job', 'metadata': { [2024-11-21 13:53:51,461] kopf.objects [DEBUG ] [tink/reboot-icadmin012-to-pxe] Patching with: {'metadata': {'annotations': {'kopf.zalando.org/last-handled-configuration': '{"spec [2024-11-21 13:53:51,469] kopf.objects [DEBUG ] [tink/icadmin012] Something has changed, but we are not interested (the essence is the same). [2024-11-21 13:53:51,469] kopf.objects [DEBUG ] [tink/icadmin012] Handling cycle is finished, waiting for new changes.

Is there any reason the second handler is never called ?

[julb]

I have the exact same issue but with batch/v1/jobs. I want to be notified when a job is completed in a @kopf.on.update handler but it never triggers. Having Something has changed, but we are not interested (the essence is the same).

[jaepetto]

@julb I don't know if it will help you, but I worked around the problem by using a timer. Instead of catching the update event, I'm checking the object every n seconds to check if the new value I'm expecting is here. If not, I simply do nothing. It's a bit old school, but it does the trick.

[julb]

Hey @jaepetto Thanks for your feedback. On my side, since I was interested in status change, I did: @kopf.on.update("batch", "v1", "jobs", field="status") Now, the handler is triggerred correctly. May this help.