It's come time to think about branching for the 0.98 release. I'm thinking about creating the branch in SVN on 12/1.
Soon thereafter over here at Intel we are planning to airlift a simian army into the EC2 cloud to put the branch through the wringer. There are 63 open JIRAs with 0.98 fix version: https://issues.apache.org/jira/browse/HBASE-8115?jql=project%20%3D%20HBASE%20AND%20fixVersion%20%3D%20%220.98.0%22%20AND%20status%20in%20(Open%2C%20%22In%20Progress%22%2C%20Reopened) That's too many. I will be moving a bunch out over the next few days. Currently there are 4 blockers: HBASE-6873 (clean up coprocessor loading failure), HBASE-9997 (Add per KV security details to HBase book), HBASE-1005 (TestVisibilityLabels fails occasionally), and HBASE-10031 (Add a section on the transparent CF encryption feature to the manual). There are 6 criticals: HBASE-5954 (add proper fsync support), HBASE-7057 (store ServerLoad in a table), HBASE-7781 (Provide the ability to run integration tests with security enabled), HBASE-8770 (deletes and puts with the same ts should be resolved according to mvcc/seqnum), HBASE-9513 (why is pe#randomseekscantest way slower in 0.96 than 0.94), and HBASE-9527 (remove all old api that takes a table name as a byte array). Given the current state of these JIRAs, I think all are likely to be moved out except HBASE-7781. During the RC process I will implement HBASE-7781 if it's not been done elsewhere (patch?) so we can get security coverage during RC integration testing. Of the remaining issues, 44 are major, 8 are minor, and 1 is trivial. Some of the major issues are follow ups to the recently committed security features. I have it on good authority these will be completed this week. Likewise, HBASE-9997 and HBASE-10031. -- Best regards, - Andy Problems worthy of attack prove their worth by hitting back. - Piet Hein (via Tom White)
