[ https://issues.apache.org/jira/browse/HBASE-11837?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jeffrey Zhong updated HBASE-11837: ---------------------------------- Attachment: HBASE-11837.patch After tracing the steps in scan code path, I found that region server does return result back through CellScanner while CoprocessorHConnection#callBlockingMethod just drops CellScanner quietly. This is "data loss" issue because client drops data on the fly. I tested with the test and it passed: {noformat} Running org.apache.phoenix.end2end.TenantSpecificTablesDDLIT Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 100.844 sec - in org.apache.phoenix.end2end.TenantSpecificTablesDDLIT {noformat} > Scanner from RegionCoprocessorEnvironment.getTable(TableName) returns no > local data > ----------------------------------------------------------------------------------- > > Key: HBASE-11837 > URL: https://issues.apache.org/jira/browse/HBASE-11837 > Project: HBase > Issue Type: Bug > Components: Coprocessors > Affects Versions: 0.98.4 > Reporter: James Taylor > Assignee: Jeffrey Zhong > Attachments: HBASE-11837.patch, PHOENIX-1208-torepro.patch > > > Might be fixed as a byproduct of the fix for HBASE-11766. > Here's what I'm seeing: > - From an endpoint coprocessor on SYSTEM.CATALOG table > - Rows exist in the table, but are all in a single region > - Scan is a new, empty scan created like this: new Scan() > - The following scanner *does* return data: > {code} > RegionScanner scanner = region.getScanner(scan); > {code} > - The following scanner *does not* return any data: > {code} > HTableInterface hTable = env.getTable(region.getTableDesc().getTableName()); > {code} > - The following scanner *does* return data: > {code} > HTablePool pool = new HTablePool (env.getConfiguration(),1); > HTableInterface hTable = pool.getTable("SYSTEM.CATALOG"); > {code} > [~jeffreyz], [~apurtell] -- This message was sent by Atlassian JIRA (v6.2#6252)