zhangbutao commented on PR #4744: URL: https://github.com/apache/hive/pull/4744#issuecomment-1875628291
@dengzhhu653 Not sure if i understand correctly. > Applying the change Hive may suffer from intermittent outage when the CBO is on(HiveServer2 connects to the old HMS but with the schema change). Do you mean that [upgrade-4.0.0-beta-1-to-4.0.0-beta-2.mysql.sql](https://github.com/apache/hive/pull/4744/files#diff-674383c98c7d2bfcd67a71401b6c14f0ba20a37d2eb11bffb767bf4e006a0a52) schema is upgraded but the hms is not? I think yes. HMS will encounter error when fetching the table&partition stats as the direct sql generated by HMS can query the underlying metadata in mysql. But i think it should not be a problem, users should know that if they upgraded the metadata schema, they also should upgrade the HMS binary package at the same. > How about leaving the redundant columns on database but HMS doesn't use these columns at all(using the multi-join proposed in this PR)? IMO, It's not a good idea that leaving the redundant columns in order for older HMS stats working. Now that you've upgrade the metadata schema, why don't you upgrade the HMS binary package? I can't think of a reason for this. Also, it seems that we never promised that the old hms can work well with the newer&upgraded metadata schema. Maybe any change on metadata schema can make old HMS occur error, not just this PR. -- 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]
