Varun Kapoor created AMBARI-646:
-----------------------------------
Summary: Design & Implement A UI Flow For Rack Information
Key: AMBARI-646
URL: https://issues.apache.org/jira/browse/AMBARI-646
Project: Ambari
Issue Type: Bug
Affects Versions: 1.0
Reporter: Varun Kapoor
Fix For: 1.0
Once https://issues.apache.org/jira/browse/AMBARI-645 is done (or even while
it's being worked on, really), we need to decide upon a first cut flow in the
Ambari UI to get rack information (from the user) about the hosts in the
cluster.
A first proposal that emanated from a 5-minute brainstorming session between
Hitesh, Arun and I was to enhance the format of the hosts file (that is
uploaded) to take in optional rack information about each host - something in
the form of:
{code}
/<rack>/<host>/
.
.
.
/<rack>/<host>/
{code}
This allows for leaving the <rack> field blank, if that information is not
known, by just making an entry of the form:
{code}
//<host>/
{code}
Alternately, we could consider deducing when there is rack information in a
host entry, by looking for '/' - if it's not present, we assume the entry is
simply of the form:
{code}
<host>
{code}
just like it is today.
Another preference I have that the implementer might want to consider is to use
a sectioned INI file format, like so:
{code}
[<rack>]
<host>
.
.
.
<host>
[<rack>]
<host>
.
.
.
<host>
{code}
Note that this just addresses getting in the rack info via the UI and storing
it in the Ambari DB - there's more work to be done (likely in a separate JIRA,
to be opened by whomever knows enough about it to actually put in a meaningful
description) in propagating this rack info to all the Hadoop components via the
Puppet layer.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
