Hi,

On 10 August 2016 at 14:11, Davide Giannella <[email protected]> wrote:

> On 09/08/2016 13:18, Ian Boston wrote:
> > Alternatively, move the indexes so that a sync property index update
> > doesn't perform a conditional change to the  global root document ? ( A
> new
> > thread would be required to discuss this if worth talking about.)
>
> I'm stubborn and maybe even slow in learning, but again I ask myself:
> why are we storing the indexes in the repository itself?
>

me also.


>
> I was not part of the original discussion around this; but frankly I
> would have expected to have the indexes stored separately from the
> repository. Let's say on the file system. Something like JR2 where it
> was even possible to delete a directory and all the indexes were
> re-generated from scratch.
>
> What do we loose if we would be moving the indexes outside of the
> repository?


Possibly the index having a revision made visible by the root revision,
however, I think there is a filter that removes later revisions from a
result set in the case of a Lucene index.

Whatever provided the index would need to be as fast or faster than Oak to
be synchronous.


> Which means each AEM node will have its own index(es).
>

Or better, sharded indexes with local indexes managed independently (each
committing to memory not disk with a WAL to deal with failures) so the cost
of indexing is parallelised and can scale horizontally... which is one step
beyond the Hybrid Index proposal.


>
> Cheers
> Davide
>
>
>
Best Regards
Ian

Reply via email to