hive-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sergey Shelukhin (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (HIVE-19631) reduce epic locking in AbstractService
Date Mon, 21 May 2018 18:29:00 GMT

     [ https://issues.apache.org/jira/browse/HIVE-19631?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Sergey Shelukhin updated HIVE-19631:
------------------------------------
    Status: Patch Available  (was: Open)

[~thejas] can you take a look?
I wonder esp. why getDelegationTokenFromMetaStore, which is remote call, is synchronized?
does it really need to be synchronized at all, and would it be enough if there's a lock that
covers only the instances of this method, without synchornizing it with everything else under
the sun?

> reduce epic locking in AbstractService
> --------------------------------------
>
>                 Key: HIVE-19631
>                 URL: https://issues.apache.org/jira/browse/HIVE-19631
>             Project: Hive
>          Issue Type: Bug
>            Reporter: Sergey Shelukhin
>            Assignee: Sergey Shelukhin
>            Priority: Major
>         Attachments: HIVE-19631.patch
>
>
> Some services have synchronized method that do lengthy remote calls, and these block
everyone trying to e.g. getHiveConf. 



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Mime
View raw message