You are viewing a plain text version of this content. The canonical link for it is here.

Posted to users@pulsar.apache.org by Adam Brenner via users <us...@pulsar.apache.org> on 2023/04/10 20:33:26 UTC

Pulsar and Tiered Storage -- Any difference in functionality?

Hi!

When we configure Pulsar to tier data (segments from Bookkeeper) at the 
namespace or topic level does this change any of the functionality 
Pulsar offers -- example: Pulsar SQL (Trino) is able to search both data 
in bookkeeper and that in tiered storage?

I understand that reading data from tiered storage might be slower 
(object storage, NFS, etc.) compared to a bookkeeper cluster but wonder 
what, if any, functionality is gained or/and lost when data is tiered.

Thanks,
-Adam

Re: Pulsar and Tiered Storage -- Any difference in functionality?

Posted by Asaf Mesika <as...@gmail.com>.

The Presto connector located in Apache Pulsar repository supports reading also offloaded segments.

> On 10 Apr 2023, at 23:33, Adam Brenner via users <us...@pulsar.apache.org> wrote:
> 
> Hi!
> 
> When we configure Pulsar to tier data (segments from Bookkeeper) at the namespace or topic level does this change any of the functionality Pulsar offers -- example: Pulsar SQL (Trino) is able to search both data in bookkeeper and that in tiered storage?
> 
> I understand that reading data from tiered storage might be slower (object storage, NFS, etc.) compared to a bookkeeper cluster but wonder what, if any, functionality is gained or/and lost when data is tiered.
> 
> Thanks,
> -Adam