[
https://issues.apache.org/jira/browse/RANGER-3703?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Binhua Hu updated RANGER-3703:
------------------------------
Description:
In the docker container or distributed nodes, the initialization
script(setup.sh) may be executed many times. Ranger controls the initialization
process through database table（x_db_version_h）. However, when the number of
database connections is too large or the network is abnormal, the
initialization script will judge abnormally, the script will trigger
reconstruction, and the data will be lost.So is it reasonable to drop tables in
script(ranger_core_db_XX.sql).
For example, the related table of Ranger has been created before, and another
docker container is started to execute the setup.sh script. However, due to
network problems, the criterion fails and the reconstruction operation is being
triggered.The initialization log is as follows，
!ranger log.png!
was:
In the docker container or distributed nodes, the initialization
script(setup.sh) may be executed many times. Ranger controls the initialization
process through database table（x_db_version_h）. However, when the number of
database connections is too large or the network is abnormal, the
initialization script will judge abnormally, the script will trigger
reconstruction, and the data will be lost.So is it reasonable to drop tables in
script(ranger_core_db_XX.sql).
The initialization log is as follows，
!ranger log.png!
The data volume in the Ranger database is as follows，
!ranger database volume.png|width=957,height=360!
> Data loss during initialization
> --------------------------------
>
> Key: RANGER-3703
> URL: https://issues.apache.org/jira/browse/RANGER-3703
> Project: Ranger
> Issue Type: Improvement
> Components: admin
> Affects Versions: 1.2.0, 2.2.0
> Reporter: Binhua Hu
> Priority: Major
> Attachments: ranger log.png
>
>
> In the docker container or distributed nodes, the initialization
> script(setup.sh) may be executed many times. Ranger controls the
> initialization process through database table（x_db_version_h）. However, when
> the number of database connections is too large or the network is abnormal,
> the initialization script will judge abnormally, the script will trigger
> reconstruction, and the data will be lost.So is it reasonable to drop tables
> in script(ranger_core_db_XX.sql).
> For example, the related table of Ranger has been created before, and another
> docker container is started to execute the setup.sh script. However, due to
> network problems, the criterion fails and the reconstruction operation is
> being triggered.The initialization log is as follows，
> !ranger log.png!
>
>
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
