frostruan commented on PR #5699: URL: https://github.com/apache/hbase/pull/5699#issuecomment-1966341246
I think there are two problems here: 1. For in-transition regions, the result of RegionLocation.getServerName() could be null. In the current PR, we just filter out these regions. With in-transition region missing, we may lost some data. 2. We first get all region servers with regions of target table, then we request each region server to get RegionMetrics. If any region is moved from this server during this period, this region will be missed too. For the first problem, I think maybe we can return something indicates that we can not know the specific data size now, so in the previous discussion, I propose introducing a new constant UNKNOWN_SIZE with value of 1 byte. For the second problem, I think maybe we'd better use a snapshot of cluster metrics to make sure we will not miss any region. What do you think ? Thanks. @aalhour -- 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]
