Closed wangxf1987 closed 2 weeks ago
Hi Xiaofei, let me try to format the content and cc interpreter owners @chaunceyjiang @XiShanYongYe-Chang
Hi @wangxf1987 Is it the kservice CRD resource that cannot be propagated or the CRD CR resource that cannot be propagated?
Have you checked the logs of karmada-controller-manager
for errors?
Hi @wangxf1987 Is it the kservice CRD resource that cannot be propagated or the CRD CR resource that cannot be propagated?
Have you checked the logs of
karmada-controller-manager
for errors?
@XiShanYongYe-Chang the prepagatation is ok, but reflect status is not sync in the resource interper. i check autoscaling in the fedcluster, it's Ready in the subcluster actuallly. I try to modify resync-period for setupController, but no effect. I think that the sync is caused by the resource not having an UPDATE EVENT.
the sync is work in here:
Hi @wangxf1987 Is it the kservice CRD resource that cannot be propagated or the CRD CR resource that cannot be propagated? Have you checked the logs of
karmada-controller-manager
for errors?@XiShanYongYe-Chang the prepagatation is ok, but reflect status is not sync in the resource interper. i check autoscaling in the fedcluster, it's Ready in the subcluster actuallly. I try to modify resync-period for setupController, but no effect. I think that the sync is caused by the resource not having an UPDATE EVENT.
I think that the sync is caused by the resource not having an UPDATE EVENT.
The work-status-controller
watch to resource changes in member clusters. Events should not be lost. If an error occurs during status collection, the status cannot be collected properly. Therefore, I want to know whether the controller logs report errors.
You can use the karmadactl interpret
command to test whether the ResourceInterpreterCustomization lua configuration you wrote is correct.
I think that the sync is caused by the resource not having an UPDATE EVENT.
The
work-status-controller
watch to resource changes in member clusters. Events should not be lost. If an error occurs during status collection, the status cannot be collected properly. Therefore, I want to know whether the controller logs report errors.
Please wait a moment, I will reproduce it again
You can use the
karmadactl interpret
command to test whether the ResourceInterpreterCustomization lua configuration you wrote is correct.
I run resource interperter of lua in the local, it's ok.
I think that the sync is caused by the resource not having an UPDATE EVENT.
The
work-status-controller
watch to resource changes in member clusters. Events should not be lost. If an error occurs during status collection, the status cannot be collected properly. Therefore, I want to know whether the controller logs report errors.
I found error logs
E0411 07:00:26.895109 1 controller.go:329] "Reconciler error" err="Operation cannot be fulfilled on kservices.fed.jd.com \"andy-benchmarkdtest35-xcp884\": the object has been modified; please apply your changes to the latest version and try again" controller="resourceBinding_status_controller" namespace="andy" name="andy-benchmarkdtest35-xcp884-kservice" reconcileID=8e05ab7d-e1f9-4214-8438-0d18fa44958b
E0411 07:00:26.895044 1 common.go:78] Failed to update resource(fed.jd.com/v1, Resource=kservices/andy/andy-benchmarkdtest35-xcp884), Error: Operation cannot be fulfilled on kservices.fed.jd.com "andy-benchmarkdtest35-xcp884": the object has been modified; please apply your changes to the latest version and try again
E0411 07:00:26.225524 1 controller.go:329] "Reconciler error" err="Operation cannot be fulfilled on kservices.fed.jd.com \"andy-benchmarkdtest50-gdqdcg\": the object has been modified; please apply your changes to the latest version and try again" controller="resourceBinding_status_controller" namespace="andy" name="andy-benchmarkdtest50-gdqdcg-kservice" reconcileID=a10309ac-1cb2-4663-a85e-a48d1e5fe1bf
E0411 07:00:26.225457 1 common.go:78] Failed to update resource(fed.jd.com/v1, Resource=kservices/andy/andy-benchmarkdtest50-gdqdcg), Error: Operation cannot be fulfilled on kservices.fed.jd.com "andy-benchmarkdtest50-gdqdcg": the object has been modified; please apply your changes to the latest version and try again
E0411 07:00:02.238893 1 controller.go:329] "Reconciler error" err="Operation cannot be fulfilled on kservices.fed.jd.com \"andy-benchmarkdtest48-24lfct\": the object has been modified; please apply your changes to the latest version and try again" controller="resourceBinding_status_controller" namespace="andy" name="andy-benchmarkdtest48-24lfct-kservice" reconcileID=ee878d49-07c7-40d4-b1b9-5eecf2731305
but all resources are synced, i understand that the retry.RetryOnConflict will be hander this error.
This error doesn't matter. The error indicates that the resource object obtained by the update has been modified.
but all resources are synced, i understand that the retry.RetryOnConflict will be hander this error.
The process of collecting resource status is as follows:
Work status -> RB/CRB status -> resource template status
How about seeing if the status of work and binding is collected correctly.
This error doesn't matter. The error indicates that the resource object obtained by the update has been modified.
but all resources are synced, i understand that the retry.RetryOnConflict will be hander this error.
The process of collecting resource status is as follows:
Work status -> RB/CRB status -> resource template status
How about seeing if the status of work and binding is collected correctly.
thx for your reply, and what role does the executation controller play?
thx for your reply, and what role does the executation controller play?
It parses the resource template from the work and creates it in the member cluster.
Hi @wangxf1987, any progress on your problem?
Hi @wangxf1987, any progress on your problem?
@XiShanYongYe-Chang
Hi @wangxf1987, any progress on your problem?
@XiShanYongYe-Chang
no probelm about sync, I performed modified kservice all night. All normally. I use the Resync for informer. https://github.com/karmada-io/karmada/pull/4825
Hi @wangxf1987 do you mean that the content described in the current issue has been resolved by modifying #4825?
Hi @wangxf1987 do you mean that the content described in the current issue has been resolved by modifying #4825?
it seems that at present, i will be continue to test.
Hi @wangxf1987, the problem didn't recur?
What do you think about this question?:
my resource interpertertation of lua as below:
Environment:
Karmada version: 1.6.1
Kubernetes version: 1.19.10
Others: