[
https://issues.apache.org/jira/browse/HIVE-18538?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
anishek updated HIVE-18538:
---------------------------
Description: Currently for bootstrap load we dynamically generate the task
DAG at runtime rather than generate the whole task graph at compile / semantic
analysis time. The number of vertex as part of this DAG is controlled via
"hive.repl.approx.max.load.tasks". Till now we only have tested with
replication with one database and providing the destination database name in
the repl load command, the ability to update the last.repl.id is dependent on
this. Given we want to move to a db regex for dumping data we have to relook at
this implementation (was: Currently for bootstrap load we dynamically generate
the task DAG at runtime rather than generate the whole task graph at compile /
semantic analysis time. The number of vertex as part of this DAG is controlled
via "hive.repl.approx.max.load.tasks". Till now we only have tested with
replication with one database and providing the destination database name in
the repl load command, the ability to update the last.repl.id is dependent on
this. Given we want to move to a db regex for dumping data we have to rethink
this feature.
)
> Updating last.repl.id on replica warehouse when dbName specified as a pattern
> -----------------------------------------------------------------------------
>
> Key: HIVE-18538
> URL: https://issues.apache.org/jira/browse/HIVE-18538
> Project: Hive
> Issue Type: Bug
> Components: HiveServer2
> Reporter: anishek
> Priority: Major
> Fix For: 3.0.0
>
>
> Currently for bootstrap load we dynamically generate the task DAG at runtime
> rather than generate the whole task graph at compile / semantic analysis
> time. The number of vertex as part of this DAG is controlled via
> "hive.repl.approx.max.load.tasks". Till now we only have tested with
> replication with one database and providing the destination database name in
> the repl load command, the ability to update the last.repl.id is dependent on
> this. Given we want to move to a db regex for dumping data we have to relook
> at this implementation
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
