You are viewing a plain text version of this content. The canonical link for it is here.
Posted to solr-user@lucene.apache.org by "Wayne L. Johnson" <wj...@familysearch.org> on 2017/09/27 18:19:44 UTC

Solr 7.0.0 -- can it use a 6.5.0 data repository (index)

I'm testing Solr 7.0.0.  When I start with an empty index, Solr comes up just fine, I can add documents and query documents.  However when I start with an already-populated set of documents (from 6.5.0), Solr will not start.  The relevant portion of the traceback seems to be:
Caused by: java.lang.NullPointerException
        at java.util.Objects.requireNonNull(Objects.java:203)
            ...
        at java.util.stream.ReferencePipeline.reduce(ReferencePipeline.java:479)
        at org.apache.solr.index.SlowCompositeReaderWrapper.<init>(SlowCompositeReaderWrapper.java:76)
        at org.apache.solr.index.SlowCompositeReaderWrapper.wrap(SlowCompositeReaderWrapper.java:57)
        at org.apache.solr.search.SolrIndexSearcher.<init>(SolrIndexSearcher.java:252)
        at org.apache.solr.core.SolrCore.openNewSearcher(SolrCore.java:2034)
        ... 12 more

In looking at the de-compiled code (SlowCompositeReaderWrapper), lines 72-77, and it appears that one or more "leaf" files doesn't have a "min-version" set.  That's a guess.  If so, does this mean Solr 7.0.0 can't read a 6.5.0 index?

Thanks

Wayne Johnson
801-240-4024
wjohnson000@ldschurch.org<ma...@ldschurch.org>
[familysearch2.JPG]


Re: Solr 7.0.0 -- can it use a 6.5.0 data repository (index)

Posted by Cassandra Targett <ca...@gmail.com>.
Regarding not finding the issue, JIRA has a problem with queries when
the user is not logged in (see also
https://jira.atlassian.com/browse/JRASERVER-38511 if you're interested
in the details). There's unfortunately not much we can do about it
besides manually edit issues to remove a security setting which gets
automatically added to issues when they are created (which I've now
done for SOLR-11406).

Your best bet in the future would be to log into JIRA before
initiating a search to be sure you aren't missing one that's "hidden"
inadvertently.

Cassandra

On Wed, Sep 27, 2017 at 1:39 PM, Wayne L. Johnson
<wj...@familysearch.org> wrote:
> First, thanks for the quick response.  Yes, it sounds like the same problem!!
>
> I did a bunch of searching before repoting the issue, I didn't come across that JIRA or I wouldn't have reported it.  My apologies for the duplication (although it is a new JIRA).
>
> Is there a good place to start searching in the future?  I'm a fairly experiences Solr user, and I don't mind slogging through Java code.
>
> Meanwhile I'll follow the JIRA so I know when it gets fixed.
>
> Thanks!!
>
> -----Original Message-----
> From: Stefan Matheis [mailto:matheis.stefan@gmail.com]
> Sent: Wednesday, September 27, 2017 12:32 PM
> To: solr-user@lucene.apache.org
> Subject: Re: Solr 7.0.0 -- can it use a 6.5.0 data repository (index)
>
> That sounds like https://urldefense.proofpoint.com/v2/url?u=https-3A__issues.apache.org_jira_browse_SOLR-2D11406&d=DwIFaQ&c=z0adcvxXWKG6LAMN6dVEqQ&r=4gLDKHTqOXldY2aQti2VNXYWPtqa1bUKE6MA9VrIJfU&m=iYU948dQo6G0tKFQUguY6SHOZNZoCOEAEv1sCf4ukcA&s=HvPPQL--s3bFtNyBdUiz1hNIqfLEVrb4Cu-HIC71dKY&e=  if i'm not mistaken?
>
> -Stefan
>
> On Sep 27, 2017 8:20 PM, "Wayne L. Johnson" <wj...@familysearch.org>
> wrote:
>
>> I’m testing Solr 7.0.0.  When I start with an empty index, Solr comes
>> up just fine, I can add documents and query documents.  However when I
>> start with an already-populated set of documents (from 6.5.0), Solr
>> will not start.  The relevant portion of the traceback seems to be:
>>
>> Caused by: java.lang.NullPointerException
>>
>>         at java.util.Objects.requireNonNull(Objects.java:203)
>>
>>             …
>>
>>         at java.util.stream.ReferencePipeline.reduce(
>> ReferencePipeline.java:479)
>>
>>         at org.apache.solr.index.SlowCompositeReaderWrapper.<init>(
>> SlowCompositeReaderWrapper.java:76)
>>
>>         at org.apache.solr.index.SlowCompositeReaderWrapper.wrap(
>> SlowCompositeReaderWrapper.java:57)
>>
>>         at org.apache.solr.search.SolrIndexSearcher.<init>(
>> SolrIndexSearcher.java:252)
>>
>>         at org.apache.solr.core.SolrCore.openNewSearcher(SolrCore.java:
>> 2034)
>>
>>         ... 12 more
>>
>>
>>
>> In looking at the de-compiled code (SlowCompositeReaderWrapper), lines
>> 72-77, and it appears that one or more “leaf” files doesn’t have a
>> “min-version” set.  That’s a guess.  If so, does this mean Solr 7.0.0
>> can’t read a 6.5.0 index?
>>
>>
>>
>> Thanks
>>
>>
>>
>> Wayne Johnson
>>
>> 801-240-4024
>>
>> wjohnson000@ldschurch.org
>>
>> [image: familysearch2.JPG]
>>
>>
>>

RE: Solr 7.0.0 -- can it use a 6.5.0 data repository (index)

Posted by "Wayne L. Johnson" <wj...@familysearch.org>.
First, thanks for the quick response.  Yes, it sounds like the same problem!!

I did a bunch of searching before repoting the issue, I didn't come across that JIRA or I wouldn't have reported it.  My apologies for the duplication (although it is a new JIRA).

Is there a good place to start searching in the future?  I'm a fairly experiences Solr user, and I don't mind slogging through Java code.

Meanwhile I'll follow the JIRA so I know when it gets fixed.

Thanks!!

-----Original Message-----
From: Stefan Matheis [mailto:matheis.stefan@gmail.com] 
Sent: Wednesday, September 27, 2017 12:32 PM
To: solr-user@lucene.apache.org
Subject: Re: Solr 7.0.0 -- can it use a 6.5.0 data repository (index)

That sounds like https://urldefense.proofpoint.com/v2/url?u=https-3A__issues.apache.org_jira_browse_SOLR-2D11406&d=DwIFaQ&c=z0adcvxXWKG6LAMN6dVEqQ&r=4gLDKHTqOXldY2aQti2VNXYWPtqa1bUKE6MA9VrIJfU&m=iYU948dQo6G0tKFQUguY6SHOZNZoCOEAEv1sCf4ukcA&s=HvPPQL--s3bFtNyBdUiz1hNIqfLEVrb4Cu-HIC71dKY&e=  if i'm not mistaken?

-Stefan

On Sep 27, 2017 8:20 PM, "Wayne L. Johnson" <wj...@familysearch.org>
wrote:

> I’m testing Solr 7.0.0.  When I start with an empty index, Solr comes 
> up just fine, I can add documents and query documents.  However when I 
> start with an already-populated set of documents (from 6.5.0), Solr 
> will not start.  The relevant portion of the traceback seems to be:
>
> Caused by: java.lang.NullPointerException
>
>         at java.util.Objects.requireNonNull(Objects.java:203)
>
>             …
>
>         at java.util.stream.ReferencePipeline.reduce(
> ReferencePipeline.java:479)
>
>         at org.apache.solr.index.SlowCompositeReaderWrapper.<init>(
> SlowCompositeReaderWrapper.java:76)
>
>         at org.apache.solr.index.SlowCompositeReaderWrapper.wrap(
> SlowCompositeReaderWrapper.java:57)
>
>         at org.apache.solr.search.SolrIndexSearcher.<init>(
> SolrIndexSearcher.java:252)
>
>         at org.apache.solr.core.SolrCore.openNewSearcher(SolrCore.java:
> 2034)
>
>         ... 12 more
>
>
>
> In looking at the de-compiled code (SlowCompositeReaderWrapper), lines 
> 72-77, and it appears that one or more “leaf” files doesn’t have a 
> “min-version” set.  That’s a guess.  If so, does this mean Solr 7.0.0 
> can’t read a 6.5.0 index?
>
>
>
> Thanks
>
>
>
> Wayne Johnson
>
> 801-240-4024
>
> wjohnson000@ldschurch.org
>
> [image: familysearch2.JPG]
>
>
>

Re: Solr 7.0.0 -- can it use a 6.5.0 data repository (index)

Posted by Stefan Matheis <ma...@gmail.com>.
That sounds like https://issues.apache.org/jira/browse/SOLR-11406 if i'm
not mistaken?

-Stefan

On Sep 27, 2017 8:20 PM, "Wayne L. Johnson" <wj...@familysearch.org>
wrote:

> I’m testing Solr 7.0.0.  When I start with an empty index, Solr comes up
> just fine, I can add documents and query documents.  However when I start
> with an already-populated set of documents (from 6.5.0), Solr will not
> start.  The relevant portion of the traceback seems to be:
>
> Caused by: java.lang.NullPointerException
>
>         at java.util.Objects.requireNonNull(Objects.java:203)
>
>             …
>
>         at java.util.stream.ReferencePipeline.reduce(
> ReferencePipeline.java:479)
>
>         at org.apache.solr.index.SlowCompositeReaderWrapper.<init>(
> SlowCompositeReaderWrapper.java:76)
>
>         at org.apache.solr.index.SlowCompositeReaderWrapper.wrap(
> SlowCompositeReaderWrapper.java:57)
>
>         at org.apache.solr.search.SolrIndexSearcher.<init>(
> SolrIndexSearcher.java:252)
>
>         at org.apache.solr.core.SolrCore.openNewSearcher(SolrCore.java:
> 2034)
>
>         ... 12 more
>
>
>
> In looking at the de-compiled code (SlowCompositeReaderWrapper), lines
> 72-77, and it appears that one or more “leaf” files doesn’t have a
> “min-version” set.  That’s a guess.  If so, does this mean Solr 7.0.0 can’t
> read a 6.5.0 index?
>
>
>
> Thanks
>
>
>
> Wayne Johnson
>
> 801-240-4024
>
> wjohnson000@ldschurch.org
>
> [image: familysearch2.JPG]
>
>
>