[
https://issues.apache.org/jira/browse/YARN-4920?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15263543#comment-15263543
]
Xuan Gong commented on YARN-4920:
---------------------------------
Thanks for the review, [~djp]
bq. It is possible for an app to match with the other one, like: abcd_011 can
also be matched as abcd_01. Isn't it? If so, we should fix this issue.
No, It will not.
bq. We shouldn't allocate memory of buffer inside of while loop which will
cause a lot of unncessary GC operations in case log file is pretty large.
Instead, we should reuse the limited size buffer.
Interesting. This exists in several different places. Instead of handling them
separately, let us fix it together after the refactory patch.
bq. I would suggest to add file's modification time to
response.header("Last-Modified", file_modification_time_truncate_to_seconds)
also. User may need to figure out which app's log could be old and which apps'
log are new which shouldn't depend on file's download time via http. However,
this is just an optional comments, see if you want to address now or later.
Good point. I think that yarn log command may have the same problem. Let us
address them later.
Attached a new patch to address other comments
> ATS/NM should support a link to dowload/get the logs in text format
> -------------------------------------------------------------------
>
> Key: YARN-4920
> URL: https://issues.apache.org/jira/browse/YARN-4920
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: yarn
> Reporter: Xuan Gong
> Assignee: Xuan Gong
> Attachments: YARN-4920.2.patch, YARN-4920.20160424.branch-2.patch
>
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
