Requesting a persistent volume in Trident returns with an EOF error
Applies to
- Trident for Kubernetes 19.07 and higher
- Trident for Openshift 19.07 and higher
Issue
The following error ends the PV creation attempt: Post "http://127.0.0.1:8000/trident/v1/backend": EOF
Additional symptoms and behavior seen:
- worker nodes have full connectivity and resolve all names correctly
/etc/resolv.conf
in worker nodes have a "search domains" directive, that works as expected- pods running on these worker nodes will inherit options and search domains from their host, and add a ndots:5 directive
- pods only resolve FQDNs, but when removing "ndots:5" or search domains inside the pod, the pod is able to resolve any name
- apart from name resolution, some destinations are unreachable (ping : no route to host)
- the kube-apiserver is slow to answer any query with kubectl
- an ONTAPI https request test to the SVM shows that it's responding normally to, for example a system-get-version API call (*)