What is the best way to set the settings for the _percolator index? I
create an index benchmark and then let elasticsearch create _percolatorupon registration of the first query. Should I be manually creating the
_percolator index myself so that I can make the settings what I want? Or
is there a way to store off the settings and have elasticsearch read them
automatically when it creates the _percolator index?
On Thu, 2012-06-07 at 07:53 -0700, shadow000fire wrote:
What is the best way to set the settings for the _percolator index? I
create an index benchmark and then let elasticsearch create
_percolator upon registration of the first query. Should I be
manually creating the _percolator index myself so that I can make the
settings what I want? Or is there a way to store off the settings and
have elasticsearch read them automatically when it creates the
_percolator index?
Easiest is probably just to create the _percolator index yourself,
before creating any percolators
clint
What settings are you after?
On Thu, Jun 7, 2012 at 4:53 PM, shadow000fire jason.scheller@gmail.comwrote:
What is the best way to set the settings for the _percolator index? I
create an index benchmark and then let elasticsearch create *_percolator
- upon registration of the first query. Should I be manually creating
the _percolator index myself so that I can make the settings what I
want? Or is there a way to store off the settings and have elasticsearch
read them automatically when it creates the _percolator index?
Whatever ones will get me sub 100 ms percolate responses with >= 100K registered queries.
I think I'm looking for how to distribute percolate requests the way that searches are distributed. But if there's a better approach then I'm looking for that.
Thanks,
Jay
On Jun 10, 2012, at 5:38 PM, Shay Banon kimchy@gmail.com wrote:
What settings are you after?
On Thu, Jun 7, 2012 at 4:53 PM, shadow000fire jason.scheller@gmail.com wrote:
What is the best way to set the settings for the _percolator index? I create an index benchmark and then let elasticsearch create _percolator upon registration of the first query. Should I be manually creating the _percolator index myself so that I can make the settings what I want? Or is there a way to store off the settings and have elasticsearch read them automatically when it creates the _percolator index?
You can have several indexes holding parts of the percolated queries, and
execute them in parallel. Its not amazing, and I hope we address this in
the upcoming release.
On Mon, Jun 11, 2012 at 4:24 AM, Jason Scheller jason.scheller@gmail.comwrote:
Whatever ones will get me sub 100 ms percolate responses with >= 100K
registered queries.I think I'm looking for how to distribute percolate requests the way that
searches are distributed. But if there's a better approach then I'm looking
for that.Thanks,
JayOn Jun 10, 2012, at 5:38 PM, Shay Banon kimchy@gmail.com wrote:
What settings are you after?
On Thu, Jun 7, 2012 at 4:53 PM, shadow000fire jason.scheller@gmail.comwrote:
What is the best way to set the settings for the _percolator index? I
create an index benchmark and then let elasticsearch create *
_percolator* upon registration of the first query. Should I be manually
creating the _percolator index myself so that I can make the settings
what I want? Or is there a way to store off the settings and have
elasticsearch read them automatically when it creates the _percolatorindex?