hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sangjin Lee (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-4958) The file localization process should allow for wildcards to reduce the application footprint in the state store
Date Fri, 10 Jun 2016 19:44:21 GMT

    [ https://issues.apache.org/jira/browse/YARN-4958?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15325157#comment-15325157
] 

Sangjin Lee commented on YARN-4958:
-----------------------------------

Sorry [~templedf] it took me a while to get back to this.

I finally got around to trying out the patch with a pseudo-distributed setup. I can confirm
that the main use cases seem to work correctly, including the case of a non-jar resource in
the staging directory. That said, I do have some high level comments as well as a couple of
minor nits.

(1)
Regarding the decision of determining the public-ness solely based on the parent directory
in the case of the wildcard, I'm wondering whether that would have any implications. It's
probably not going to be common, but it is possible that the directory is public but there
may be files that are not readable by others. Again, it's hard to imagine why one would do
this, but if they did, would it cause a security issue on localization or a localization failure?
Should we chalk that up to an unsupported setting? To be fair, I can see this being an issue
if a directory was specified (not a wildcard), too. In that sense, we could say this is a
manifestation of an existing issue... Thoughts?

(2)
With {{ContainerExecutor.java}}, what happens if the wildcarded directory has further nested
directories? It appears we're symlinking only at the immediate child level. I suspect it would
work correctly, but wanted to double check.

(3)
Were you able to test it in the local job mode?

(4) {{ClientDistributedCacheManager.java}}
- l.303: change {{System.out.println()}} to a logger logging statement

(5) {{DistributedCache.java}}
- l.295: typo: "it's" -> "its"

Finally, since most of the changes are really in MAPREDUCE, perhaps this JIRA should be moved
to the MAPREDUCE project. What do you think? If we really want to follow the rules to the
letter, we would need to create separate JIRAs for all projects involved (HADOOP, YARN, and
MAPREDUCE). I'd like to hear what you think. On a related note, you may want to drop the changes
to {{Path.java}} if you can help it.


> The file localization process should allow for wildcards to reduce the application footprint
in the state store
> ---------------------------------------------------------------------------------------------------------------
>
>                 Key: YARN-4958
>                 URL: https://issues.apache.org/jira/browse/YARN-4958
>             Project: Hadoop YARN
>          Issue Type: Improvement
>          Components: nodemanager
>    Affects Versions: 2.8.0
>            Reporter: Daniel Templeton
>            Assignee: Daniel Templeton
>            Priority: Critical
>         Attachments: YARN-4958.001.patch, YARN-4958.002.patch, YARN-4958.003.patch
>
>
> When using the -libjars option to add classes to the classpath, every library so added
is explicitly listed in the {{ContainerLaunchContext}}'s local resources even though they're
all uploaded to the same directory in HDFS.  When using tools like Crunch without an uber
JAR or when trying to take advantage of the shared cache, the number of libraries can be quite
large.  We've seen many cases where we had to turn down the max number of applications to
prevent ZK from running out of heap because of the size of the state store entries.
> Rather than listing all files independently, this JIRA proposes to have the NM allow
wildcards in the resource localization paths.  Specifically, we propose to allow a path to
have a final component (name) set to "*", which is interpreted by the NM as "download the
full directory and link to every file in it from the job's working directory."  This behavior
is the same as the current behavior when using -libjars, but avoids explicitly listing every
file.
> This JIRA does not attempt to provide more general purpose wildcards, such as "\*.jar"
or "file\*", as having multiple entries for a single directory presents numerous logistical
issues.
> This JIRA also does not attempt to integrate with the shared cache.  That work will be
left to a future JIRA.  Specifically, this JIRA only applies when a full directory is uploaded.
 Currently the shared cache does not handle directory uploads.
> This JIRA proposes to allow for wildcards both in the internal processing of the -libjars
switch and in paths added through the {{Job}} and {{DistributedCache}} classes.
> The proposed approach is to treat a path, "dir/\*", as "dir" for purposes of all file
verification and localization.  In the final step, the NM will query the localized directory
to get a list of the files in "dir" such that each can be linked from the job's working directory.
 Since $PWD/\* is always included on the classpath, all JAR files in "dir" will be in the
classpath.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: yarn-issues-help@hadoop.apache.org


Mime
View raw message