[jira] [Commented] (SENTRY-1643) AutoIncrement ChangeID of MSentryPermChange/MSentryPathChange may be error-prone

["Lei (Eddy) Xu (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SENTRY-1643?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15954397#comment-15954397 ] 

Lei (Eddy) Xu commented on SENTRY-1643:
---------------------------------------

I believe that this is a concurrency issue when do the parallel perm/path changes insertions in large transactions. 

It is similar to this: http://stackoverflow.com/questions/1716515/mysql-autoincrement-column-increases-by-10-problem. 

It'd be difficult for a ORM to make the primary keys strictly consecutive.  [~hahao], [~akolb] is there a reason that requires the change ID be consecutive? Would monotone increasing change IDs become sufficient? I do feel that not relying consecutive values can be more robust.

> AutoIncrement ChangeID of MSentryPermChange/MSentryPathChange may be error-prone
> --------------------------------------------------------------------------------
>
>                 Key: SENTRY-1643
>                 URL: https://issues.apache.org/jira/browse/SENTRY-1643
>             Project: Sentry
>          Issue Type: Sub-task
>          Components: Hdfs Plugin
>    Affects Versions: sentry-ha-redesign
>            Reporter: Hao Hao
>            Assignee: Lei (Eddy) Xu
>            Priority: Critical
>             Fix For: sentry-ha-redesign
>
>
> In MSentryPermChange/MSentryPathChange table, the changeID field is auto-increment. 
> {noformat}
> <field name="changeID" primary-key="true" value-strategy="increment">
>   <column name="CHANGE_ID" jdbc-type="BIGINT" allows-null="false"/>
> </field>{noformat}
> However, found through the integration unit test TestHDFSIntegration, the value does not seem to be correctly auto increased. e.g Instead of increasing by one for each new entry, it increased by some unexpected number.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)