[
https://issues.apache.org/jira/browse/HDFS-12161?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Siyao Meng updated HDFS-12161:
------------------------------
Description:
After running distcp as user 'usertest' which is a hdfs superuser, the
destination folder ownership is not preserved.
e.g.
{code:bash}
hadoop distcp -pugpaxt -update -skipcrccheck /tmp/usertest
/tmp/new_user/usertest_copy
{code}
After distcp is executed, the parent folder has ownership with 'usertest' :
{code:bash}
drwxr-xr-x - usertest supergroup 0 2017-07-13 22:09
/tmp/new_user/usertest_copy
{code}
and the actual files copied by distcp preserved the permissions:
{code:bash}
drwxr-xr-x - hive hive 0 2017-07-13 22:09
/tmp/new_user/usertest_copy/dir1
-rw-r--r-- 1 hdfs hive 287 2017-07-13 22:09
/tmp/new_user/usertest_copy/test1.txt
{code}
The distcp options (-pugpaxt) does not apply to the destination parent
directory specified in the distcp command e.g. "/tmp/new_user/usertest_copy".
Can we document this in https://hadoop.apache.org/docs/r1.2.1/distcp2.html
under Command Line Options
The destination folder location needs to be pre-created with correct
ownership/permissions before using distcp. The preserve options do not apply to
parent folder.
was:
After running distcp as user 'usertest' which is a hdfs superuser, the
destination folder ownership is not preserved.
e.g.
{code:bash}
hadoop distcp -pugpaxt -update -skipcrccheck /tmp/usertest
/tmp/new_user/usertest_copy
{code}
After distcp is executed, the parent folder has ownership with 'usertest' :
{code:bash}
drwxr-xr-x - usertest supergroup 0 2017-07-13 22:09
/tmp/new_user/usertest_copy
{code}
and the actual files copied by distcp preserved the permissions:
{code:bash}
drwxr-xr-x - hive hive 0 2017-07-13 22:09
/tmp/new_user/usertest_copy/dir1
-rw-r--r-- 1 hdfs hive 287 2017-07-13 22:09
/tmp/new_user/usertest_copy/test1.txt
{code}
The distcp options (-pugpaxt) does not apply to the destination parent
directory specified in the distcp command e.g. "/tmp/new_user/usertest_copy"
Can we document this in https://hadoop.apache.org/docs/r1.2.1/distcp2.html
under Command Line Options
The destination folder location needs to be pre-created with correct
ownership/permissions before using distcp. The preserve options do not apply to
parent folder.
> Distcp does not preserve ownership in destination parent folder
> ----------------------------------------------------------------
>
> Key: HDFS-12161
> URL: https://issues.apache.org/jira/browse/HDFS-12161
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs
> Affects Versions: 2.6.0
> Reporter: Sailesh Patel
> Assignee: Siyao Meng
> Priority: Minor
>
> After running distcp as user 'usertest' which is a hdfs superuser, the
> destination folder ownership is not preserved.
> e.g.
> {code:bash}
> hadoop distcp -pugpaxt -update -skipcrccheck /tmp/usertest
> /tmp/new_user/usertest_copy
> {code}
> After distcp is executed, the parent folder has ownership with 'usertest' :
> {code:bash}
> drwxr-xr-x - usertest supergroup 0 2017-07-13 22:09
> /tmp/new_user/usertest_copy
> {code}
> and the actual files copied by distcp preserved the permissions:
> {code:bash}
> drwxr-xr-x - hive hive 0 2017-07-13 22:09
> /tmp/new_user/usertest_copy/dir1
> -rw-r--r-- 1 hdfs hive 287 2017-07-13 22:09
> /tmp/new_user/usertest_copy/test1.txt
> {code}
> The distcp options (-pugpaxt) does not apply to the destination parent
> directory specified in the distcp command e.g. "/tmp/new_user/usertest_copy".
> Can we document this in https://hadoop.apache.org/docs/r1.2.1/distcp2.html
> under Command Line Options
> The destination folder location needs to be pre-created with correct
> ownership/permissions before using distcp. The preserve options do not apply
> to parent folder.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
