Installed Couchbase on minikube deployed on a linux (Ubuntu) Virtual Machine
Installation seems to work but
- From the web UI the loading of the sample buckets (Travel, beer, etc) fail (they are created and deleted immediately)
- thinking about memory problems I tried to increase RAM for DATA from 256 the a bigger value. It does not work
- worst problem. After rebooting minikube (minikube stop/start) or rebooting the Ubuntu VM, the 3 pods of Couchbase go in ERROR state and Couchbase doesn’t work anymore
ANyboy has already met this problem ? I hae already installed couchbase directly on a VM (without using containers) and it works
So I think there are some problems delivering Couchbase in a minikube environment
I’m guessing you are using the operator:
- The operator manages buckets so will delete any that you haven’t configured in the custom resource. Depending on the version you are using see spec.disableBucketManagement or spec.buckets.managed to disable this.
- Again memory management is controlled exclusively by the operator see spec.cluster.dataServiceMemoryQuota
- You haven’t deployed your cluster with persistent storage, so there is nothing to recover from and all your data has been lost. Use persistent storage!
Hopefully that should get you up and running!
I am quite new to minicube
so I used istructions found here
I don’t if they istructiona follow your advices …
Those quick-start instructions only deploy the absolute minimum required to demonstrate the Operator. That will not provide any HA features.
To make your cluster able to survive a restart of minkube you can learn about persistent volumes here https://docs.couchbase.com/operator/2.0/concept-persistent-volumes.html and there is a guide how to configure it here https://docs.couchbase.com/operator/2.0/howto-persistent-volumes.html.
You can learn more about memory allocation here https://docs.couchbase.com/operator/2.0/concept-memory-allocation.html if you want to increase your memory quotas.
Finally this configuration attribute https://docs.couchbase.com/operator/2.0/reference-couchbasecluster.html#spec-buckets-managed will allow you to turn off bucket management. That will allow you to load up any of the samples that server provides.
If you need any more help just ask, also any suggestions to make the documentation better would be appreciated.
Thank to your advices I resolved the 2 problems of the bucket creation from UI and Memory allocation
About the persistant volumes it is not clear to me if the indication reported in the page
Could be implemented after the deployment (and how)
or I have to restart from scratch modifying the file provided in the installatio guide page
Thank you for help
Persistent Volumes can be added after deployment. An in-place upgrade of the Couchbase Pods will occurr:
Modifying any field in the spec.servers object that affects the underlying pod will trigger an upgrade, as pods are immutable...