All,
Need advice on ES prod configuration,Our use case is very straightforward
and simple.
Data is updated only once per week and mostly the data would be searched.
We have 3 nodes and planning to configure them as follows.
1 nodes with data true and master false, 2 nodes with data and master true.
we have a web server acting as proxy on top of the 3 nodes, hence did not
choose the node client route.
also if i bulk insert data should i be inserting into data=true
,master=false or the other 2 nodes
is this a good idea?
Another question, in Test we were running 2 boxes with master and data
true(default)
we bulk inserted node 1 and dont see data replica in node2, why would that
happen, i set the number_replicas to 1 in the config.
You'd be better off setting them all as master and data true, it'll reduce
the possibility for a split brain situation to occur.
In that instance, and with your proxy, it doesn't really matter too much
where you bulk index data.
Did you specify replicas when you created the index? If you didn't then it
should replicate based on your setting.
All,
Need advice on ES prod configuration,Our use case is very straightforward
and simple.
Data is updated only once per week and mostly the data would be searched.
We have 3 nodes and planning to configure them as follows.
1 nodes with data true and master false, 2 nodes with data and master true.
we have a web server acting as proxy on top of the 3 nodes, hence did not
choose the node client route.
also if i bulk insert data should i be inserting into data=true
,master=false or the other 2 nodes
is this a good idea?
Another question, in Test we were running 2 boxes with master and data
true(default)
we bulk inserted node 1 and dont see data replica in node2, why would that
happen, i set the number_replicas to 1 in the config.
Thank you, as i finished the post and was reading documentation, it made
sense to make all of them master and data as true.
i haven't specifically mentioned replica when i bulk index, if i specify
what would be the syntax, so that i can go back and double check.
On Monday, March 23, 2015 at 6:58:38 PM UTC-5, Mark Walkom wrote:
You'd be better off setting them all as master and data true, it'll reduce
the possibility for a split brain situation to occur.
In that instance, and with your proxy, it doesn't really matter too much
where you bulk index data.
Did you specify replicas when you created the index? If you didn't then it
should replicate based on your setting.
On 24 March 2015 at 09:27, TB <txin...@gmail.com <javascript:>> wrote:
All,
Need advice on ES prod configuration,Our use case is very straightforward
and simple.
Data is updated only once per week and mostly the data would be searched.
We have 3 nodes and planning to configure them as follows.
1 nodes with data true and master false, 2 nodes with data and master
true.
we have a web server acting as proxy on top of the 3 nodes, hence did not
choose the node client route.
also if i bulk insert data should i be inserting into data=true
,master=false or the other 2 nodes
is this a good idea?
Another question, in Test we were running 2 boxes with master and data
true(default)
we bulk inserted node 1 and dont see data replica in node2, why would
that happen, i set the number_replicas to 1 in the config.
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.