Tejaskriya commented on PR #10: URL: https://github.com/apache/ozone-helm-charts/pull/10#issuecomment-3308204860
@pyttel that makes sense, I understand the difficulties due to the clashes between the design of Kubernetes and Raft. But here we are assuming that the volume (pv/pvc) in which the OM or SCM store their data can be moved around. In a case where we are storing the data locally on the host machine, this would not be possible. Say a node goes bad, then kubernetes would choose a different node for the OM or SCM to come up. But on this node, the metadata isn't present => we lose the data. This is where HA is needed. Hence this decision should be based on the type of storage the user configures. Given we choose local storage, I think an option of HA would be needed. What StorageClass have you been using for your deployments? A kubernetes operator would work as well, but I haven't done enough research on that to give a strong opinion. If I am not mistaken, that would take a bigger effort than the current PR. Maybe eventually if we can have both options that would be better? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
