You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hive.apache.org by "Sergey Shelukhin (JIRA)" <ji...@apache.org> on 2018/05/21 23:22:00 UTC

[jira] [Updated] (HIVE-19642) add cache dump when LLAP cache is full on branch-2 (and fix retry issues)

     [ https://issues.apache.org/jira/browse/HIVE-19642?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Sergey Shelukhin updated HIVE-19642:
------------------------------------
    Summary: add cache dump when LLAP cache is full on branch-2 (and fix retry issues)  (was: add cache dump when LLAP cache has issues on branch-2 (and fix retry issues))

> add cache dump when LLAP cache is full on branch-2 (and fix retry issues)
> -------------------------------------------------------------------------
>
>                 Key: HIVE-19642
>                 URL: https://issues.apache.org/jira/browse/HIVE-19642
>             Project: Hive
>          Issue Type: Bug
>            Reporter: Sergey Shelukhin
>            Assignee: Sergey Shelukhin
>            Priority: Major
>         Attachments: HIVE-19642-branch-2.patch
>
>
> LLAP cache doesn't have defragmentation on branch-2 so it retries when failing to evict.
> There are a couple of issues with retries (e.g. sleep is not applied properly and time can go negative), also it would be helpful to make use of existing cache dump logic used in LLAP IO view to dump log state into the log in such cases.



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