Apache9 commented on pull request #3447: URL: https://github.com/apache/hbase/pull/3447#issuecomment-878319143
> @Apache9 @virajjasani Sorry, I was out for the holiday. I understand that this is a special region and it does not have a corresponding table/namespace created for it. We do have an internal patch/feature that makes use of the system table designation (see [HBASE-18477](https://issues.apache.org/jira/browse/HBASE-18477)). We block writes/updates for user tables, but not system tables ([HBASE-18775](https://issues.apache.org/jira/browse/HBASE-18775)). Without this flag being set, writes to the master local region will be blocked with our feature. Originally the system table status was stored in the HRegionInfo, but now only the Table knows if it is a system table or not. > > I can add a hack/hardcode a check for the master local region. However, it still seems like this is technically a system region? In that case, why would we not denote it as a system region? It certainly is not in the user's domain, it is an internal HBase region. > > I'm fine with not adding this to the reserved NS list/restricting creating the NS, but to me it still seems like it is a bug that this "table" (really region) isn't being considered a system table. It is not a table actually... Neither a user table nor a system table... So I suggest we should have a way to bypass the newly added logic, and we set this flag or config when constructing the master local region. Thanks. -- 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]
