Closed howespt closed 4 years ago
it might be worth re-deploying with the KubernetesContainerFactory (see https://github.com/apache/openwhisk-deploy-kube/blob/master/docs/configurationChoices.md#invoker-container-factory) and see if that helps. The fact that the invoker is doing a docker rm
right after the inspect suggests that it thinks the container is unhealthy in some way.
Can confirm adding
invoker:
containerFactory:
impl: "kubernetes"
to the cluster config yaml fixed the issue. Thanks!
I wonder where to add the commands shown above
Might be related to #545
Managed to install openwhisk and all pods/services/etc running as expected.
Trying various options for running an action results in the same error. e.g. invoking a whisk.system sample action
wsk -i action invoke /whisk.system/samples/greeting --result
results inerror: Unable to invoke action 'samples/greeting': The connection failed, or timed out. (HTTP status code 500)
Looking at the activation I see
Digging into the logs everything seems fine except I see the following line from the controller pod:
[2019-11-14T21:03:22.200Z] [INFO] [#tid_HviouueynrGRdZZUNNR61Ng1Niw4ps0W] [ShardingContainerPoolBalancer] received completion ack for '9b2f121039c74fe5af121039c7afe55c', system error=true
The invoker pod logs show:
From here I'm having difficulty tracing the error. Any ideas on where to proceed from here?
Additional context which may relate to #545 If I continue to try to invoke the actions then I put all of the invokers into an unhealthy state and get an unable to invoke error. Looking at the logs the health checks seem to work but the invokers remain in "unhealthy" status.