Hi,
We are giving elasticsearch a test drive on our production environment as a
replacement for couchdb-lucene and during initial testing on VM things
looked great, but, now we have moved to real data and we are seeing some
really bad performance in it pulling initial data from our couchdb.
According to bigdesk we are seeing only 30 indexes per second and we have
over 200 millions docs in our couchdb so you can imagine how long this is
going to take. And we get way more then 30 docs created per second in our
couchdb. You can also see that the box is not resource starved as we are
using one of the larger AWS instances. And looking at our couchdb instance
it is also not being taxed. What type of tuning/configuration do we need
to do to get elasticsearch running at 100% performance or do we just need
to wait till it catches up and then we should expect it to be working at
optimal performaces.
Thanks
Zuhaib
--