[
https://issues.apache.org/jira/browse/HBASE-4284?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13093282#comment-13093282
]
Wing Yew Poon commented on HBASE-4284:
--------------------------------------
Related to this, the usage printed when importtsv is run without expected
arguments is misleading:
{noformat}
In order to prepare data for a bulk data load, pass the option:
-Dimporttsv.bulk.output=/path/for/output
Note: if you do not use this option, then the target table must
already exist in HBase
{noformat}
This gives the user the impression that if they use this option, the target
table need not already exist in HBase. I find that not to be the case.
The usage statement should be revised to remove that note.
> document permissions that need to be set on importtsv output before
> completebulkload
> ------------------------------------------------------------------------------------
>
> Key: HBASE-4284
> URL: https://issues.apache.org/jira/browse/HBASE-4284
> Project: HBase
> Issue Type: Improvement
> Components: documentation, mapreduce
> Affects Versions: 0.90.4
> Reporter: Wing Yew Poon
>
> I am using HBase 0.94 from CDH3u1.
> After running importtsv using the -Dimporttsv.bulk.output=<output dir>
> option, I find that completebulkload fails due to hbase not having write
> permissions on the contents of the output dir that importtsv wrote. I have to
> manually set write permissions on these contents before I can run
> completebulkload successfully.
> Ideally, I should not have to do that (set the permissions manually). Given
> that I do, this should at least be documented as a limitation of the
> importtsv utility.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
