[ https://issues.apache.org/jira/browse/YARN-6608?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16035458#comment-16035458 ]
Carlo Curino commented on YARN-6608: ------------------------------------ [~yufeigu] The question you ask if very reasonable, yet non-trivial to answer. My main concern is that there are some patches (e.g., HDFS-10266) which we do not fully import---that patch specifically is an HDFS patch, which only modifies a .json file in SLS. This was one of the concerns we discussed with [~leftnoteasy] and [~chris.douglas] early on, as we might be backporting "parts of a patch". This might be enough for the SLS to work properly, though might make later backports of patchs a little more clumsy (you will have to exclude the portion of the patch that touches SLS). Backporting all full patches is likely overkill/does not make sense, but it is needed for SLS to work. [~leftnoteasy]'s point to me was that this raw approach is better than living with the choppy/partial implementation of SLS we have in branch-2. Moreover, I see from the last yetus that there is some JDK 1.7 issues to look at (and I don't have much time for this right now, so help is welcome here). What do you guys think about this? Is it worth doing like this? > Backport all SLS improvements from trunk to branch-2 > ---------------------------------------------------- > > Key: YARN-6608 > URL: https://issues.apache.org/jira/browse/YARN-6608 > Project: Hadoop YARN > Issue Type: Improvement > Affects Versions: 2.9.0 > Reporter: Carlo Curino > Assignee: Carlo Curino > Attachments: YARN-6608-branch-2.v0.patch, > YARN-6608-branch-2.v1.patch, YARN-6608-branch-2.v2.patch > > > The SLS has received lots of attention in trunk, but only some of it made it > back to branch-2. This patch is a "raw" fork-lift of the trunk development > from hadoop-tools/hadoop-sls. -- This message was sent by Atlassian JIRA (v6.3.15#6346) --------------------------------------------------------------------- To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org