[bitnami/elasticsearch] in Dualstack IPv6 & SVC mismatch #24531
Description
Name and Version
bitnami/elasticsearch
What architecture are you using?
amd64
What steps will reproduce the bug?
Install bitnami/elasticsearch chart in k8s with DualStack IP configuration.
I have an IP DualStack k8s configuration. Interestingly, elasticsearch in DualStack config fallback to using IPv6 and testing service connectivity via IPv6, but by default, it created an IPv4-only service.
So I needed to modify the zammad-elasticsearch & zammad-elasticsearch-master-hl services by modifying spec to these:
Spec:
ipFamilies:
- IPv4
- IPv6
ipFamilyPolicy: PreferDualStackAlso, to make this work, I needed to remove livenessProbe & readinessProbe for the elasticsearch StetfulSet:
/spec/template/spec/containers/0/readinessProbe
/spec/template/spec/containers/0/livenessProbe
I discovered this using the official zammad helm chart 10.3.4, which uses elastic search as a sub-chart and opened a ticket for them.
Are you using any custom parameters or values?
No response
What is the expected behavior?
In the Dualstack configuration, if services are created using IPv4, then pods should communicate via that IPv4.
If there are two types of IP, then even if one is reachable, elastic search should be working.
Or create services with IPv4 & IPv6 by default for DualStack k8s.
What do you see instead?
Pods are trying to communicate via IPv6 to services with IPv4 and obviously failing.
Additional information
No response