[ https://issues.apache.org/jira/browse/KAFKA-3250?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15195538#comment-15195538 ]
ASF GitHub Bot commented on KAFKA-3250: --------------------------------------- GitHub user granthenke opened a pull request: https://github.com/apache/kafka/pull/1075 KAFKA-3250: release tarball is unnecessarily large due to duplicate l… …ibraries This ensures duplicates are not copied in the distribution without rewriting all of the tar'ing logic. A larger improvement could be made to the packaging code, but that should be tracked by another jira. You can merge this pull request into a Git repository by running: $ git pull https://github.com/granthenke/kafka libs-duplicates Alternatively you can review and apply these changes as the patch at: https://github.com/apache/kafka/pull/1075.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #1075 ---- commit 8cdbf18fb5b751e0fc922d405643c152daaef4d1 Author: Grant Henke <granthe...@gmail.com> Date: 2016-03-15T15:53:43Z KAFKA-3250: release tarball is unnecessarily large due to duplicate libraries This ensures duplicates are not copied in the distribution without rewriting all of the tar'ing logic. A larger improvement could be made to the packaging code, but that should be tracked by another jira. ---- > release tarball is unnecessarily large due to duplicate libraries > ----------------------------------------------------------------- > > Key: KAFKA-3250 > URL: https://issues.apache.org/jira/browse/KAFKA-3250 > Project: Kafka > Issue Type: Bug > Affects Versions: 0.9.0.1 > Reporter: Gwen Shapira > Assignee: Grant Henke > Fix For: 0.10.0.0 > > > Between 0.8.2.2 and 0.9.0, our release tarballs grew from 17M to 34M. We > thought it is just due to new libraries and dependencies. But: > 1. If you untar Kafka into a directory and check the directory size (du -sh), > it is around 28M, smaller than the tarball. Recompressing give you 25M > tarball. > 2. If you list the original tar contents and grep for "snappy", you see it 4 > times in the tarball. > Clearly we are creating a tarball with duplicates (and we didn't before). > I think its due to how we are generating the tarball from core but pull in > other projects into libs/ directory with their dependencies (which overlap). > We need to find out how to sort it out (possibly with excludes). -- This message was sent by Atlassian JIRA (v6.3.4#6332)