Mailing-List: contact dev-help@accumulo.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@accumulo.apache.org
Content-Type: multipart/alternative;
 boundary="===============4830327131931285537=="
MIME-Version: 1.0
Subject: Re: Review Request 17860: Review of changes that remove the necessity
 for
 ACCUMULO_LOG_HOST
From: keith@deenlo.com
To: "accumulo" <dev@accumulo.apache.org>, "Josh Elser" <josh.elser@gmail.com>,
 keith@deenlo.com
Date: Sat, 08 Feb 2014 00:48:33 -0000
Message-ID: <20140208004833.22694.42339@reviews.apache.org>
Auto-Submitted: auto-generated
Sender: noreply@reviews.apache.org
References: <20140207223056.22694.68539@reviews.apache.org>
In-Reply-To: <20140207223056.22694.68539@reviews.apache.org>
Reply-To: keith@deenlo.com

--===============4830327131931285537==
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/17860/#review33988
-----------------------------------------------------------


server/src/main/java/org/apache/accumulo/server/watcher/MonitorLog4jWatcher.java
<https://reviews.apache.org/r/17860/#comment63888>

    If getData() throws NoNodeException (or returns null? not sure if that will happen) should probably gracefully handle this and stop trying to send log data wherever it used to be sending it.


- kturner


On Feb. 7, 2014, 10:30 p.m., Josh Elser wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/17860/
> -----------------------------------------------------------
> 
> (Updated Feb. 7, 2014, 10:30 p.m.)
> 
> 
> Review request for accumulo.
> 
> 
> Bugs: ACCUMULO-2334
>     https://issues.apache.org/jira/browse/ACCUMULO-2334
> 
> 
> Repository: accumulo
> 
> 
> Description
> -------
> 
> Removes the necessity to compute and set ACCUMULO_LOG_HOST. We can accomplish this by having the monitor advertise a host in addition to just the port and have the other processes place a watcher on that node in zookeeper.
> 
> This has some nice benefits to simplifying the code, removing the annoying (IMO) ACCUMULO_LOG_HOST variable, and, when running the monitor on a random port, the other services can auto-discover without having to update configuration files and restart the services. The auto-discovery also has benefit if the monitor has to be moved to a different host (or multiple monitors are being run with "failover")
> 
> 
> Diffs
> -----
> 
>   conf/examples/1GB/native-standalone/accumulo-env.sh aa4a1d0 
>   conf/examples/1GB/standalone/accumulo-env.sh 1707f3d 
>   conf/examples/2GB/native-standalone/accumulo-env.sh ef74ca7 
>   conf/examples/2GB/standalone/accumulo-env.sh 75014c2 
>   conf/examples/3GB/native-standalone/accumulo-env.sh ae0da11 
>   conf/examples/3GB/standalone/accumulo-env.sh 7edd938 
>   conf/examples/512MB/native-standalone/accumulo-env.sh 749a678 
>   conf/examples/512MB/standalone/accumulo-env.sh 9beb059 
>   core/src/main/java/org/apache/accumulo/core/Constants.java 9bb3419 
>   core/src/main/java/org/apache/accumulo/core/util/AddressUtil.java af9a1a6 
>   server/src/main/java/org/apache/accumulo/server/Accumulo.java f4da33b 
>   server/src/main/java/org/apache/accumulo/server/monitor/LogService.java 10ef9e4 
>   server/src/main/java/org/apache/accumulo/server/monitor/Monitor.java dcb80fd 
>   server/src/main/java/org/apache/accumulo/server/watcher/MonitorLog4jWatcher.java PRE-CREATION 
> 
> Diff: https://reviews.apache.org/r/17860/diff/
> 
> 
> Testing
> -------
> 
> Local testing so far. Configure the monitor to use a random log4j port, start an instance. Then, test log forwarding works. Restart the monitor, check that the tserver saw the update, and that log forwarding still works even though the monitor is now listening on a different port for log4j.
> 
> 
> Thanks,
> 
> Josh Elser
> 
>


--===============4830327131931285537==--