Mailing-List: contact dev-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@hbase.apache.org
Received-SPF: neutral (athena.apache.org: local policy)
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Subject: Re: Review Request: Patch for hbase-2599 for 0.20.5 inclusion
From: "Todd Lipcon" <todd@cloudera.com>
To: "Todd Lipcon" <todd@cloudera.com>, jiraposter@review.hbase.org,
	dev@hbase.apache.org, stack@duboce.net
Date: Wed, 26 May 2010 21:01:35 -0000
Message-ID: <20100526210135.20565.84826@ip-10-250-10-165.ec2.internal>
In-Reply-To: <20100525230424.7069.84964@ip-10-250-10-165.ec2.internal>
References: <20100525230424.7069.84964@ip-10-250-10-165.ec2.internal>


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://review.hbase.org/r/88/#review81
-----------------------------------------------------------

Ship it!


this seems reasonable, but only way we'll know is testing on various system=
s with different kinds of DNS setups. Can we get those who have had issues =
to try this and make sure it fixes, before we release it?


branches/0.20/src/java/org/apache/hadoop/hbase/ClusterStatus.java
<http://review.hbase.org/r/88/#comment409>

    I don't like that this function is called getServerNames and returns ho=
stnameports.


branches/0.20/src/test/org/apache/hadoop/hbase/TestServerInfo.java
<http://review.hbase.org/r/88/#comment410>

    using the junit4 expected thing here doesn't quite work, cuz we don't k=
now that it didn't throw on line 34, right?


- Todd


On 2010-05-25 16:04:24, stack wrote:
> =

> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://review.hbase.org/r/88/
> -----------------------------------------------------------
> =

> (Updated 2010-05-25 16:04:24)
> =

> =

> Review request for hbase.
> =

> =

> Summary
> -------
> =

> Version of hbase-2599 for 0.20 branch for 0.20.5.
> =

> It cuts DNS lookups from HServerInfo and uses the regionservers name ever=
ywhere instead of ip sometimes and a lookedup hostname at other times.  Als=
o puts hostname:port into .META. rather than IP.  Removing DNS lookups make=
s it so there is no possibility for disagreement over hostname if master ge=
ts one answer for an IP and the regionserver has another.
> =

> Other notes:
> =

> It replaces variable servername with hostnameandport so its clear whats g=
oing on.
> =

> Does some cleanup in HServerInfo renaming data member 'name' as 'hostname=
' and then purges DNS caching and lookups from this class.  It removes some=
 unnecessary setters and does some javadoc fix ups.
> =

> Removes the dumb serverAddressToServerInfo from ServerManager.
> =

> =

> This addresses bug hbase-2599.
> =

> =

> Diffs
> -----
> =

>   branches/0.20/src/java/org/apache/hadoop/hbase/ClusterStatus.java 94821=
8 =

>   branches/0.20/src/java/org/apache/hadoop/hbase/HServerInfo.java 948218 =

>   branches/0.20/src/java/org/apache/hadoop/hbase/master/BaseScanner.java =
948218 =

>   branches/0.20/src/java/org/apache/hadoop/hbase/master/HMaster.java 9482=
18 =

>   branches/0.20/src/java/org/apache/hadoop/hbase/master/ProcessRegionOpen=
.java 948218 =

>   branches/0.20/src/java/org/apache/hadoop/hbase/master/ServerManager.jav=
a 948218 =

>   branches/0.20/src/java/org/apache/hadoop/hbase/regionserver/HRegionServ=
er.java 948218 =

>   branches/0.20/src/test/org/apache/hadoop/hbase/TestServerInfo.java PRE-=
CREATION =

>   branches/0.20/src/webapps/master/table.jsp 948218 =

> =

> Diff: http://review.hbase.org/r/88/diff
> =

> =

> Testing
> -------
> =

> Doing now.
> =

> =

> Thanks,
> =

> stack
> =

>