[
https://issues.apache.org/jira/browse/HBASE-4931?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14047367#comment-14047367
]
Misty Stanley-Jones commented on HBASE-4931:
--------------------------------------------
bq. The rs.impl and rs.class do not default to ReplicationXYZ. The command line
in this jira is just an example of what you can do. and as far as I know we
don't have those ReplicationXYZ classes in the codebase.
OK, I took that out. Thanks for clearing it up.
bq. the <command>export</command> referenced in the Import, should be with the
first letter uppercase.
Fixed.
bq. At the end "For performance consider the following general options
scanner.caching=100 speculative.execution=false"
bq. speculative execution should be always off and not for performance reason
but to avoid to write data twice which may produce the wrong result.
bq. For the caching maybe say something like >= 100, or some explanation like
an higher value uses more memory but does less round trip to the server which
may improve performance (something like that)
I made this modification. Please double-check. If there is a parameter that
should always be false, why don't we just set it to false in the code?
bq. then there is another reference to the ReplicationRegionInterface and
ReplicationRegionServer which I'm not able to find anywhere in our code base
Removed that one. I'm not even sure why it was there as it was an empty
paragraph.
> CopyTable instructions could be improved.
> -----------------------------------------
>
> Key: HBASE-4931
> URL: https://issues.apache.org/jira/browse/HBASE-4931
> Project: HBase
> Issue Type: Bug
> Components: documentation, mapreduce
> Affects Versions: 0.90.4, 0.92.0
> Reporter: Jonathan Hsieh
> Assignee: Misty Stanley-Jones
> Attachments: HBASE-4931.patch
>
>
> The book and the usage instructions could be improved to include more
> details, things caveats and to better explain usage.
> One example in particular, could be updated to refer to
> ReplicationRegionInterface and ReplicationRegionServer in thier current
> locations (o.a.h.h.client.replication and o.a.h.h.replication.regionserver),
> and better explain why one would use particular arguments.
> {code}
> $ bin/hbase org.apache.hadoop.hbase.mapreduce.CopyTable
> --rs.class=org.apache.hadoop.hbase.ipc.ReplicationRegionInterface
> --rs.impl=org.apache.hadoop.hbase.regionserver.replication.ReplicationRegionServer
> --starttime=1265875194289 --endtime=1265878794289
> --peer.adr=server1,server2,server3:2181:/hbase TestTable
> {code}
--
This message was sent by Atlassian JIRA
(v6.2#6252)
