-
Notifications
You must be signed in to change notification settings - Fork 137
New issue
Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? # to your account
Topics hono.command_internal.*
are not being cleaned up on Docker versions >20.10
#3537
Comments
@calohmn would mind taking a look? |
The solution mentioned in the Stackoverflow link above - obtaining the container id from But, the general issue here is, that Example case on an AKS cluster, using containerdContainer id from Line in
Running
This issue is also mentioned in
There are initiatives to have a universal mechanism for obtaining the container ID, see
For the time being, I think the most general kind of solution in Hono would be to use the Kubernetes API client in the protocol adapter code, querying the adapter pod container status entry to get the container id from there. |
For the Kafka client id, including the container id isn't strictly necessary. It could also be some other id that is unique for all containers of a pod (maybe containing a timestamp for better readability). Suppose we keep the location of The alternative would be refactoring the method to create the Kafka client id, using some other id there. Then we could move the So, a question here would be if we want to keep @sophokles73 WDYT? |
I've created #3545 for that refactoring. In a subsequent PR, I would add the new K8s-API based |
My understanding of the Kafka client ID is that it is used to determine membership in a consumer group, right?
I guess using a simple UUID could also work here, because the consumer groups always consist of only one single consumer (a container in a pod), or am I mistaken? |
Yes, I agree. |
No, the In terms of ids in association with a consumer group, there is the
Yes, I've used a UUID (portion) in connection with the pod name in #3545. |
Signed-off-by: Carsten Lohmann <carsten.lohmann@bosch.io>
Signed-off-by: Carsten Lohmann <carsten.lohmann@bosch.io>
The
KubernetesContainerUtil.getContainerId()
method doesn't find the containerId since Docker version 20.10, see:https://stackoverflow.com/questions/68816329/how-to-get-docker-container-id-from-within-the-container-with-cgroup-v2
As a consequence the adapter instance id will be as for non-k8s, which will eventually result in
hono.command_internal.*
topics are not cleaned up when the adapter goes down.The text was updated successfully, but these errors were encountered: