Mailing-List: contact hdfs-dev-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Reply-To: hdfs-dev@hadoop.apache.org
Received-SPF: pass (athena.apache.org: domain of eli@cloudera.com designates
 209.85.216.48 as permitted sender)
MIME-Version: 1.0
In-Reply-To: <D59DF28F-3861-4F8A-80EC-CDFB4EAA5C90@Holsman.NET>
References: <D59DF28F-3861-4F8A-80EC-CDFB4EAA5C90@Holsman.NET>
Date: Fri, 28 Jan 2011 10:58:04 -0800
Message-ID: <AANLkTimCuDfDmXiecG9_y+ZKndCY2u=JjCK+S3LKEiw9@mail.gmail.com>
Subject: Re: That append patch in 0.20
From: Eli Collins <eli@cloudera.com>
To: hdfs-dev@hadoop.apache.org
Content-Type: text/plain; charset=ISO-8859-1

Which patch are you referring to?

There's a whole branch to support sync on 20:
http://svn.apache.org/repos/asf/hadoop/common/branches/branch-0.20-append

In this branch dfs.support.append defaults to true (see HDFS-1254),
this configuration has been tested by quite a few people.

Thanks,
Eli

On Fri, Jan 28, 2011 at 10:49 AM, Ian Holsman <hadoop@holsman.net> wrote:
> Hey guys.
>
> I was speaking to some people a couple of nights ago, and we came up with an idea about it (which I'm sure has been mentioned before so apologies beforehand).
>
> could we make append a configuration option. either at build or at run time.
>
> that way we could let alot of the hbase users run off the same codebase and not have them patch it themselves.
>
> sure.. we will still have the same testing issues, and debates on if it is stable or properly tested.. we can put big warning messages in the documentation on when people should be using it.. but at least they could get a regular build
> and run it.
>
> so.. the question to the experts.. can the append patch be made into a configuration option such that if people don't have it enabled it won't affect them.
>
> Thanks for listening
>
> --ian