[
https://issues.apache.org/jira/browse/BIGTOP-1488?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14174802#comment-14174802
]
stanley shi commented on BIGTOP-1488:
-------------------------------------
Here's my plan:
1. implement a custom function and put at
"bigtop_toolchain/lib/puppet/parser/functions/nearest_apache_mirror.rb", it
will return the nearest apache mirror by parsing the json response from the URL
"https://www.apache.org/dyn/closer.cgi?as_json=1";;
2. in the puppet code, use the custom function to replace the apache mirror
prefixes;
But I have a questions that cannot be decided:
how to parsing the http/json response from the apache site:
a. the ideal way would be to use ruby json gem; but this will need more
dependency to be pulled in;
b. another way is to parse the json string as simple string and extract the
content from it;
I am not sure which is the better way;
> bigtop_toolchain hardcode the apache mirror sites for downloading maven
> -----------------------------------------------------------------------
>
> Key: BIGTOP-1488
> URL: https://issues.apache.org/jira/browse/BIGTOP-1488
> Project: Bigtop
> Issue Type: Improvement
> Components: build, general
> Affects Versions: 0.8.0
> Reporter: stanley shi
>
> In the bigtop_toolchain module, the puppet modules use hardcoded apache
> mirror sites to download component tar balls(for example, use
> ftp://mirror.reverse.net/pub/apache/maven/maven-3/3.0.5/binaries/apache-maven-3.0.5-bin.tar.gz
> to download maven).
> In some cases the hardcoded mirror maybe very slow, causing the installation
> to be very slow or even fail.
> We can use the "www.apache.org/dyn/closer.cgi" to choose apache mirrors
> automatically and faster the installatin.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
