Hi Shawn,

Thanks for your reply.

The core is used to test the solr setup by creating a trial collection. But
other than that every collection created are exactly the same as existing
one. I don't think this would create a problem, please brief me if it is
otherwise.

Thanks and regards,
Bhuvaneshwar Venkatraman

On Mon, Sep 16, 2019, 9:59 PM Shawn Heisey <apa...@elyograg.org> wrote:

> On 9/16/2019 10:18 AM, Bhuvanesh wrote:
> > Recently I created a cloud Solr 8.1.1 with zookeeper similar to cloud
> Solr
> > 6.6.2 which is in use. All configurations and schema files are exactly
> > alike, but when I try to index the same documents Solr throws *cannot
> > change field "FIELD_NAME" from* *index
> options=DOCS_AND_FREQS_AND_POSITIONS
> > to inconsistent index options=DOCS* for a specific field which is of type
> > *string*. It is a required field so cannot be omitted.
> >
> > For another Collection in the same core(Solr 8.1.1), Solr throws *cannot
> > change docValues type from SORTED_NUMERIC to SORTED for field
> > "ANOTHER_FIELD_NAME"* to the field of type *string.*
> >
> > This is not existing solr upgrade, Solr 8.1.1 setup is completely new
> from
> > scratch and I am simply indexing the same data which we are using now
> with
> > solr 6.6.2
>
> Those errors indicate that there is an existing index that was built
> with a different schema, and the current index is not compatible with
> the schema you have now.
>
> You're saying that you're indexing from scratch, but the error messages
> are indicating otherwise.
>
> It's not possible to use an index from Solr 6 in Solr 8, and the errors
> would be entirely different if you tried.
>
> This problem looks like you indexed some stuff in version 8, then
> changed the schema, and are trying to index more.  But your problem
> description seems to contradict this theory, so I'm not sure what to think.
>
> The solution to these errors is to entirely delete all of the
> $CORE/data/index directories from each index core related to the
> collection, restart Solr or reload the collection, and reindex again
> from scratch.
>
> Thanks,
> Shawn
>

Reply via email to