Closed khooz closed 5 months ago
I have deleted the contents of the PV and restarted the pod. It created a new instance and is now running. How ever I have lost my data. It was OK with this test environment but I think an autonomous upgrade path should be incorporated into images. Also digging through the alert logs I have noticed that domain name is predefined. Is there a way to join or initialize a DB to a domain so it could be monitored through cloud control?
@khooz which domain name are you referring to here? and what is it pre-defined as?
I have deleted the contents of the PV and restarted the pod. It created a new instance and is now running. How ever I have lost my data. It was OK with this test environment but I think an autonomous upgrade path should be incorporated into images.
If the database wasn't configured to use a persistent volume for its data, dropping/recreating the pod deletes the data. This is true of any database, not just Oracle. :)
Thanks @oraclesean
If the database wasn't configured to use a persistent volume for its data, dropping/recreating the pod deletes the data. This is true of any database, not just Oracle. :)
I have manually setup the PV to be retained. as you can see here:
---
apiVersion: v1
kind: PersistentVolume
metadata:
name: oracledb-vol
spec:
capacity:
storage: 100Gi
accessModes:
- ReadWriteOnce
- ReadWriteMany
persistentVolumeReclaimPolicy: Retain
hostPath:
path: "/run/desktop/mnt/host/d/dev/data/oradata"
type: Directory
storageClassName: hostpath
volumeMode: Filesystem
---
persistence:
volumeName: oracledb-vol
size: 100Gi
storageClass: hostpath
## oci-bv applies to OCI block volumes. Use "standard" storageClass for dynamic provisioning in Minikube. Update as appropriate for other cloud service providers
accessMode: ReadWriteMany
It was working fine with image using 23c as brand (which is not available in the repository now). I have deleted and re-applied multiple times using the same image and PV before. I can do it now with 23ai too, but it seems there were no interoperability or upgrade path available.
This is expected. There is no upgrade path for Free DB releases. Upgrade and patching are only supported with Enterprise edition.
Since the rebranding from 23c to 23ai, the older image has been removed and a new one made available. I had initialized a single-instance DB using the old image. Unfortunately I used the
latest
tag. When I started k8s on docker desktop after the rebrand, I noticed the new image has been pulled but it won't. I used kubectl describe and log commands and found out the following statements:Here's my setup: