When we were using old style replication, I did have the suggester lexicon
replicated along with other config files, and I think I triggered a suggester 
build
on replication or maybe commit (which happens with every replication).
I remember it being kind of fussy to set up. You might want to set up an extra
downstream machine to play with until you get it right.

wunder
Walter Underwood
[email protected]
http://observer.wunderwood.org/  (my blog)

> On Mar 2, 2023, at 10:42 AM, gnandre <[email protected]> wrote:
> 
> Thanks! I am using non-cloud mode at the moment. So, there is no way to
> just index it to the index node and get it replicated to the search nodes?
> Do I have to index to each search node?
> 
> Do you know why the suggester indexing does not follow the usual search
> indexing model?
> 
> On Thu, Mar 2, 2023, 12:22 PM Walter Underwood <[email protected]>
> wrote:
> 
>> You need to send a build request to each node. I used to have some code to
>> dig out the nodes from a cluster status, then send a build to each one, but
>> I think that is marooned at my previous company. It isn’t super hard, just
>> dig it out of the JSON.
>> 
>> wunder
>> Walter Underwood
>> [email protected]
>> http://observer.wunderwood.org/  (my blog)
>> 
>>> On Mar 2, 2023, at 9:03 AM, gnandre <[email protected]> wrote:
>>> 
>>> Can anybody please answer this? Many thanks in advance!
>>> 
>>> On Wed, Feb 16, 2022 at 12:52 AM gnandre <[email protected]>
>> wrote:
>>> 
>>>> Is there a way to get suggester index replicated to all search nodes
>> from
>>>> index node? Do I need to build suggester index for each search node
>>>> separately?
>>>> 
>> 
>> 

Reply via email to