As a follow up to this, I'm attempting to apply the patches required to implement AvatarNode (HDFS-966 and HDFS-976) but am running in to issues. Entirely possible that this is due to my inexperience with this process. Per the instructions on the Apache "how to contribute" page I've run $patch -p0 < patch.file within my base Hadoop directory I'd like to patch, but keep getting Hunk failures. In each case the java file being modified doesn't exist and the hunks get thrown in a .rej file.
I also ran in to problems during the javadoc build portion of attempting the following basic build, but that could be due to missing dependancies on my system (for development). I did download and specify -Dforrest.home and -Djava5.home. ant -Djavac.args="-Xlint -Xmaxwarns 1000" clean test tar -John On 2/16/11 10:05 AM, "Joseph Coleman" <[email protected]> wrote: >Thanks for the link that was a very interesting article and I am looking >at going with that approach and carve and NFS volume from my NetApp filer >for the transactions logs between the primary and backup node. The >question is any idea on how large I need to make the NFS volume? Also >were I can grab the package from for this as well to download and test? > > >Thanks again > >From: Ted Dunning <[email protected]<mailto:[email protected]>> >Date: Wed, 16 Feb 2011 09:17:15 -0600 >To: "[email protected]<mailto:[email protected]>" ><[email protected]<mailto:[email protected]>> >Cc: Joseph Coleman ><[email protected]<mailto:[email protected]>> >Subject: Re: Master Node question HDFS > >Search for [Avatar node hadoop]. > >http://hadoopblog.blogspot.com/2010/02/hadoop-namenode-high-availability.h >tml > >On Wed, Feb 16, 2011 at 7:12 AM, Joseph Coleman ><[email protected]<mailto:[email protected]>> >wrote: >I know I cannot have more than one master node for HDFS but what if I was >to create two of them and put a Kemps Loadbalance in front of them and >point to my load balancer for my stuff would something like that work? If >not what are others doing to safe guard themselves incase their master >node kicks the bucket a standby ready to go maybe? > >
