lucene-solr-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Erick Erickson <erickerick...@gmail.com>
Subject Re: Moving to solrcloud from single instance
Date Mon, 12 Aug 2019 19:58:38 GMT
Unless you expect your index to grow, as long performance is satisfactory there’s no reason
to shard. _Replicate_ perhaps if you need to sustain a higher QPS.

Here’s a sizing blog I wrote a long time ago, but it still pertains. The short form is for
you to load test one of your machines and find out how many docs you can put on it before
it falls over. _Then_ decide whether you need to shard.

And by “performance is satisfactory”, I mean the time it takes to serve up a query. If
you need to serve more queries, simply add more replicas (i.e. have a single-shard collection).
Each replica has the entire index in that case, so if 1 machine can serve 30 QPS, replicating
twice will let you serve 90 QPS .

If you do decide to shard, two things will happen. First, some operations aren’t well supported
when you shard, group.func to name one.

Second, you’ll introduce a certain amount of overhead (balanced against each shard doing
less work to be sure).

SolrCloud (in the one-shard, replicated case) will give you some good stuff, HA/DR, failover,
expandability, etc. so I’m not discouraging moving to that. Just don’t shard etc. until
you know you need to ;)

Best,
Erick

> On Aug 12, 2019, at 3:44 PM, David Hastings <hastings.recursive@gmail.com> wrote:
> 
> I actually never had a problem with the index being larger than the memory
> for a standalone instance, but the entire index is on an SSD at least one
> my end
> 
> On Mon, Aug 12, 2019 at 3:43 PM Erie Data Systems <eriedatasys@gmail.com>
> wrote:
> 
>> I am starting the planning stages of moving from a single instance of solr
>> 8 to a solrcloud implementation.
>> 
>> Currently I have a 148GB index on a single dedicated server w 96gb ram @ 16
>> cores /2.4ghz ea. + SSD disk. The search is fast but obviously the index
>> size is greater than the physical memory, which to my understanding is not
>> a good thing.
>> 
>> I have a lot of experience with single instance but none with solrcloud. I
>> have 3 machines (other than my main 1) with the exact same hardware 96gb *
>> 3 essentially which should be plenty.
>> 
>> My issue is that im not sure where to go to learn how to set this up, how
>> many shards, how many replicas, etc and would rather hire somebody or
>> something (detailed video or document)  to guide me through the process,
>> and make decisions along the way...For example I think a shard is a piece
>> of the index... but I dont even know how to decide how many replicas or
>> what they are .....
>> 
>> Thanks everyone.
>> -Craig
>> 


Mime
View raw message