prateekm commented on PR #1669: URL: https://github.com/apache/samza/pull/1669#issuecomment-1573843322
@asautins For context, the purpose of adding file metadata to the snapshot during upload is to be able to check if we need to restore the file during a restart (e.g. local file may be stale, or for a different instance of a job). In practice, the job should have a stable uname / group name in the cluster, and local state directories should be isolated between job instances on each host, so removing the uname/group check here is quite reasonable. Although removing it from the metadata while maintaining rollback compatibility is not trivial. For the latest commit, would prefer keeping FileMetadata stateless (i.e. not use static fields). It may be created concurrently for several files. IIRC it is also serialized/deserialized as JSON Changing it to maintain uid/gid alone will not work since they can be different across hosts in a cluster for the same user / group. -- 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]
