[
https://issues.apache.org/jira/browse/PHOENIX-4797?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16524732#comment-16524732
]
ASF GitHub Bot commented on PHOENIX-4797:
-----------------------------------------
GitHub user 492066199 opened a pull request:
https://github.com/apache/phoenix/pull/306
PHOENIX-4797 file not found or file exist exception when create glob…
when create global index use -snaopshot option in mapreduce task
multi mapper use the same restore dir to work with snapshot.
1,mapper will remove the file after mapper task as the other mapper will
read the file
2,mapper will create the file when begin mapper task as the other mapper
will create the file too
the 1 will cause the file not found and the 2 will case the file is exists
I change this to let mappers use different dirs to work with snapshot,after
this path,the -snapshot option can works well .
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/492066199/phoenix master
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/phoenix/pull/306.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #306
----
commit 2eebc1af80af1bad6074ce36659e9c861a68db83
Author: 492066199 <492066199@...>
Date: 2018-06-27T07:52:42Z
PHOENIX-4797 file not found or file exist exception when create global
index use -snaopshot option
----
> file not found or file exist exception when create global index use
> -snaopshot option
> -------------------------------------------------------------------------------------
>
> Key: PHOENIX-4797
> URL: https://issues.apache.org/jira/browse/PHOENIX-4797
> Project: Phoenix
> Issue Type: Bug
> Affects Versions: 4.13.2-cdh5.11.2
> Reporter: sailingYang
> Priority: Major
>
> when use indextool with -snapshot option and if the mapreduce create multi
> mapper.this will cause the hdfs file not found or hdfs file exist
> exception。finally the mapreduce task must be failed. because the mapper use
> the same restore work dir.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
