I am currently running 4 nodes, 3 data nodes with master=false, and 1
master node, with data=false. When making changes to my
default-mapping.json file, I am uncertain if I have just not done something
right, or if I need to go put this same file on all four hosts. My setup
currently does have this file defined on each node, so I don't know if that
could actually cause problems if they get out of sync.
So I have been testing by removing a default TTL, and creating a new test
index. It would appear that I need to remove the same line from the
default-mapping.json on all my nodes.
So now I have to wonder if there is a better way? Having multiple nodes
with this same file means I could easily see consistency issues if I don't
keep them all exactly the same. If I didn't put the file on all the data
nodes, would this not be an issue? I think I originally tried that and
found that just the master having the file wasn't working for me.
On Tuesday, February 18, 2014 9:49:16 AM UTC-5, Brian wrote:
I am currently running 4 nodes, 3 data nodes with master=false, and 1
master node, with data=false. When making changes to my
default-mapping.json file, I am uncertain if I have just not done something
right, or if I need to go put this same file on all four hosts. My setup
currently does have this file defined on each node, so I don't know if that
could actually cause problems if they get out of sync.
You might be interested in Index Templates, wherein you can alternatively
define this default mapping dynamically at runtime to an entire cluster.
This way, you don't have to maintain any files.
So when done this way I can put it on the master and not all the slave
nodes. THANK YOU! I knew there had to be something better than what I was
doing. Plus this looks much more flexible. I saw this page before, but I
didn't realize it for the entire cluster where as the default-mapping is
per node.
On Tuesday, February 18, 2014 2:15:35 PM UTC-5, Binh Ly wrote:
You might be interested in Index Templates, wherein you can alternatively
define this default mapping dynamically at runtime to an entire cluster.
This way, you don't have to maintain any files.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.