TDH disk toll

[Laura Morales <la...@mail.com>]

Does anybody know how demanding is Jena/Fuseki on the hard disk? I mean in terms of I/O operations, is it going to kill my hard disk or is it optimized somehow?
Is there any benchmark for Fuseki vs disk I/O?

Re: TDH disk toll

["A. Soroka" <aj...@virginia.edu>]

Nothing that you wouldn't look at for any other server application. The size of the dataset may not make as big a difference as the character of your queries (how much scanning are they doing, are you using expansive property paths, that sort of thing). 

---
A. Soroka
The University of Virginia Library

> On Apr 6, 2017, at 7:34 AM, Laura Morales <la...@mail.com> wrote:
> 
>> It caches as much as it can so should not thrash the disk unless you
>> have a very heavy update load (in which case, you really want the disk
>> copy to be updated safely).
> 
> In the case of a read-only dataset (few 10s of GBs), is there anything in particular that I should keep an eye on, or is it going to be fine?

Re: TDH disk toll

[Laura Morales <la...@mail.com>]

> It caches as much as it can so should not thrash the disk unless you
> have a very heavy update load (in which case, you really want the disk
> copy to be updated safely).

In the case of a read-only dataset (few 10s of GBs), is there anything in particular that I should keep an eye on, or is it going to be fine?

Re: TDH disk toll

[Andy Seaborne <an...@apache.org>]

On 06/04/17 01:29, Laura Morales wrote:
> Does anybody know how demanding is Jena/Fuseki on the hard disk? I mean in terms of I/O operations, is it going to kill my hard disk or is it optimized somehow?

It caches as much as it can so should not thrash the disk unless you 
have a very heavy update load (in which case, you really want the disk 
copy to be updated safely).

> Is there any benchmark for Fuseki vs disk I/O?

Not that I know of.

     Andy