I dont know the meaning in_memory=false in context of phoenix. In context of 
hbase, that entire data will be kept in memory( not just block cache). I m 
guessing, Phoenix is also doing same thing.
By data size on disk, you can roughly estimate memory footprint. For example, 
You can load one index into memory by setting in_memory flag and see how much 
memory it took. Then u can estimate memory usage for other index on basis of 
your experiment.

Sent from my iPhone

> On Mar 22, 2016, at 9:05 AM, Sumit Nigam <[email protected]> wrote:
> 
> Thank you Anil.
> 
> Any inputs on what does IN_MEMORY = false do at index creation time?
> 
> I get the separate table part. But not sure I understood why would memory 
> footprint be equal to data size. Would this secondary index reside entirely 
> in memory? I guess only parts of it would reside on block cache, right?
> 
> Thanks for such a prompt reply.
> Sumit
> 
> 
> 
> From: anil gupta <[email protected]>
> To: "[email protected]" <[email protected]>; Sumit Nigam 
> <[email protected]> 
> Sent: Tuesday, March 22, 2016 9:11 PM
> Subject: Re: Secondary index memory footprint
> 
> Global indexes are stored in a separate hbase table. So, you can estimate 
> memory footprint by looking at the data size of that index currently.
> 
> HTH,
> Anil Gupta
> 
> On Tue, Mar 22, 2016 at 7:19 AM, Sumit Nigam <[email protected]> wrote:
> Hi,
> 
> I am trying to estimate what (if any) are the implications of accumulating 
> data in phoenix secondary index. I have a secondary index on 3 columns and 
> would like to know if anyone has an idea of how to estimate memory footprint 
> of secondary index (if any) based on number of entries in data table. The 
> index is global and not local.
> 
> Plus, what is the impact of setting IN_MEMORY=true when index is created? 
> 
> Thanks,
> Sumit
> 
> 
> 
> -- 
> Thanks & Regards,
> Anil Gupta
> 
> 

Reply via email to