SAP Hana Express on Kubernetes doesn't start after removing and recreating resources

[lcasavola]

Hello, I was playing with this tutorial : https://developers.sap.com/tutorials/gnext-deploy-kubernetes-sqlpad.html

Since I didn't realize it focused on cloud k8s cluster I didn't consider that LoadBalancer method would have not worked out in my environment . Any way both hxe and sqlpad started on a single pod but I screwed it up and decided to reinstall it from scratch.

Unfortunately since then I was not able to let the hxe-container to start again; it starts but returns these errrors:

======== Starting HANA container run script ========
Started at: Mon Sep 13 14:36:29 UTC 2021
Script parameters: --agree-to-sap-license --dont-check-system --passwords-url file:///hana/hxeconfig/password.json --dont-exit-on-error
HANA version: 2.00.054.00.1611906357
WARNING: Unsupported Linux kernel version 5.4. Supported are 3.12 and 4.4+.
Setting exit trap to keep the container alive on error ...
ERROR: Unknown host 'hxe-54f5d57895-qpztr'. The container must be started with one of these host names:
ERROR: hxe-54f5d57895-q6mnp

In this case hxe-54f5d57895-q6mnp was the name of the previous created pod before the removal of all the k8s resources.

Than, I started hxe with option --dont-exit-on-error to be able to inspect the cointaner and I found this folder:

/hana/mounts/system/hosts/hxe-54f5d57895-q6mnp

which causes the error reported. More precisly the error is generated by the execution of hdbenv.sh shell script.

For same reason /hana/mounts/system contains that folder recovered from some repository I was not be able to individuate.

Rgds

Luca Casavola

[lcasavola]

Hello, I have un update. I guess I solved my issue. The problem was that when i deleted the permanent volume with the commands:

 kubectl delete  pvc hxe-pvc,  delete  pv persistent-vol-hxe 

I didn't realize that the persistent volume's contents ( /data/hxe_pv) would not be deleted and also such a content was generated in the target node k8s chose to deploy the pod I was initially no aware of.

It was suffice to delete the k8s resources and the /data/hxe_pv folder to solve the problem and recreate hana express from scratch.

Hope you find these notes useful.

Rgds

Luca