neils-dev commented on PR #4222: URL: https://github.com/apache/ozone/pull/4222#issuecomment-1414762234
> I think it would be better to return early from `processNodeReport`, ideally only in `ReconNodeManager`. Thanks @adoroszlai. I looked at the occurrence more closely. This problem arises when recon is restart after in service with datanodes registered with it. On restart, the datanode looks up the nodes it recorded in its table and proceeds to register each node with its stored `datanodeDetails` and _null_ `nodeReport`. https://github.com/apache/ozone/blob/9fba172b3f20987dbe5c7cf0bc3c712549b171d5/hadoop-ozone/recon/src/main/java/org/apache/hadoop/ozone/recon/scm/ReconNodeManager.java#L109 It looks like the least intrusive way to avoid the NPE is an early check and return within the `SCMNodeManager.processNodeReport`. Thoughts? -- 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]
