hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sunil G (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (YARN-4418) AM Resource Limit per partition can be updated to ResourceUsage as well
Date Mon, 07 Dec 2015 16:41:11 GMT

     [ https://issues.apache.org/jira/browse/YARN-4418?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel

Sunil G updated YARN-4418:
    Attachment: 0002-YARN-4418.patch

Uploading a newer version of patching after looking into test failures.

One point to note:
AMLimit is now a part of ResourceUsage. ResourceUsage was updated when containers were allocated
to it, else there wont be any entries added to the ResourceUsage. However, AMLimit may slightly
vary here as it need not have to wait for any real resourceUsage to get it calculated. It
can take queue/partitionCapacity to get the capacity.

In first patch, I tried to calculate/update AMLimit for all accessible labels. I was slightly
in half mind here, due to some unwanted resource usage updates here came also with this (0
for all other enums in ResourceUsage) . So I changed to update AMLimit for those labels which
are now Used/Added in ResourceUsage.

[~leftnoteasy] pls help to share your point of view here so that we can finalize on this.

> AM Resource Limit per partition can be updated to ResourceUsage as well
> -----------------------------------------------------------------------
>                 Key: YARN-4418
>                 URL: https://issues.apache.org/jira/browse/YARN-4418
>             Project: Hadoop YARN
>          Issue Type: Bug
>          Components: resourcemanager
>    Affects Versions: 2.7.1
>            Reporter: Sunil G
>            Assignee: Sunil G
>         Attachments: 0001-YARN-4418.patch, 0002-YARN-4418.patch
> AMResourceLimit is now extended to all partitions after YARN-3216. Its also better to
track this ResourceLimit in existing {{ResourceUsage}} so that REST framework can be benefited
to avail this information easily. 

This message was sent by Atlassian JIRA

View raw message