[
https://issues.apache.org/jira/browse/HADOOP-9639?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13680847#comment-13680847
]
Sangjin Lee commented on HADOOP-9639:
-------------------------------------
Thanks for your comment, Chris. Yes, it looks like YARN can definitely support
what I have in mind. But it seems that it would be more a building block to get
to what I want than the complete story (correct me if I'm wrong).
I am coming specifically from a map-reduce perspective. I'd like to be able to
address concerns like submitting a mapreduce job but having the job jar and the
libjars safely cached in the shared manner, the location of these jars
referenced as part of the job classpath, adding/expiring cached jars in HDFS
and locally on the nodes as well, etc.
It seems like the distributed cache does pretty much most of that. And if I'm
not mistaken, it leverages the resource localization you speak of on the
cluster side as the building block. It's only that it's limiting sharing to a
per-job basis.
So to support this for all map-reduce jobs, it looks to me that changes to
things like JobSubmitter and YARNRunner are needed. Please let me know if I'm
way off. Thanks!
> truly shared cache for jars (jobjar/libjar)
> -------------------------------------------
>
> Key: HADOOP-9639
> URL: https://issues.apache.org/jira/browse/HADOOP-9639
> Project: Hadoop Common
> Issue Type: New Feature
> Components: filecache
> Affects Versions: 2.0.4-alpha
> Reporter: Sangjin Lee
>
> Currently there is the distributed cache that enables you to cache jars and
> files so that attempts from the same job can reuse them. However, sharing is
> limited with the distributed cache because it is normally on a per-job basis.
> On a large cluster, sometimes copying of jobjars and libjars becomes so
> prevalent that it consumes a large portion of the network bandwidth, not to
> speak of defeating the purpose of "bringing compute to where data is". This
> is wasteful because in most cases code doesn't change much across many jobs.
> I'd like to propose and discuss feasibility of introducing a truly shared
> cache so that multiple jobs from multiple users can share and cache jars.
> This JIRA is to open the discussion.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
