We have elasticsearch embedded in our product.
Most customers are asking for cluster settings ensuring high availability,
In a scenario with 4 servers, they want to make sure that if 3 of the 4
nodes goes down, the last one should be able to handle requests (and also
What is the best way to accomplish this? Query performance is usually not a
big issue, since the amount of data is limited to, say, 300.000 documents.
Also, there are not many write-operations done except when doing a manual
reindex of all documents (usually just once)
A possible solution is to run with 1 shard / 3 replicas, and set the
write_consistency to "one". This will ensure that all nodes always has the
data necessary to serve and also be able to write to index. This does
implicate that its not possible for one node to be accessible without
having contact with the other nodes.
For other customers, they may have 12 nodes, 6 nodes in two different
locations. The main priority here will also be that "as few as possible"
nodes needs be running at the same, and still work. Also, here - a scenario
where one of the locations is down while all nodes in the other locations
will be available.
Any thought around this?
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to firstname.lastname@example.org.
For more options, visit https://groups.google.com/groups/opt_out.