I shall try to explain the implementation.

Ideally the backup store should keep all fields of all uncommitted docs.

For committed docs, it must store all the non-stored fields and any
field which is a destination of copyField.

The idea of fetching all the stored fields from DB is not right I
guess . We cannot do it because the DB is available only in the master
(It is not replicated).

The schema in the description is a bit dated . instead of the field
'COMMITTED' it is called STATUS. It is an enum with values

COMITTED = 0;
UNCOMMITTED = 1;
UNCOMMITTED_MARKED_FOR_DELETE = 2
COMMITTED_MARKED_FOR_DELETE = 3

If before committing a document is overwritten multiple times we just
need to keep one version of UNCOMMITTED doc right?

The other important change is SOLR-810. It is an attempt to reduce the
datastore size.

All the field names are stored in another table . These are the
KNOWN_STRING collection .

Please ask if anymore details is required

--Noble

On Tue, Dec 2, 2008 at 10:55 PM, Yonik Seeley <[EMAIL PROTECTED]> wrote:
> On Tue, Dec 2, 2008 at 12:08 PM, Shalin Shekhar Mangar
> <[EMAIL PROTECTED]> wrote:
>> HSQLDB has a limit of upto 8GB of data. In Solr, you might want to go beyond
>> that without a commit.
>
> Ah, so you only use the DB to store the uncommitted docs so you can
> quickly reference their fields?
>
> If that's all, we might be able to get that functionality from Lucene
> (IndexWriter would need to support getting the latest stored fields by
> term at least).
>
> The other option is to actually store the stored fields permanently in
> the DB... and get them from the DB instead of Lucene when requested.
> That's the road I thought you were going down.
>
> -Yonik
>
>
>> On Tue, Dec 2, 2008 at 10:33 PM, Dawid Weiss
>> <[EMAIL PROTECTED]>wrote:
>>
>>>
>>> Isn't HSQLDB an option? Its performance ranges a lot depending on the
>>> volume of data and queries, but otherwise the license looks BSDish.
>>>
>>> http://hsqldb.org/web/hsqlLicense.html
>>>
>>> Dawid
>>>
>>
>>
>>
>> --
>> Regards,
>> Shalin Shekhar Mangar.
>>
>



-- 
--Noble Paul
  • Re: Can we use Berkley... Grant Ingersoll
    • Re: Can we use Be... Noble Paul നോബിള്‍ नोब्ळ्
      • Re: Can we us... Yonik Seeley
        • Re: Can w... Andrzej Bialecki
          • Re: C... Yonik Seeley
            • ... Jason Rutherglen
              • ... Andrzej Bialecki
              • ... Dawid Weiss
              • ... Shalin Shekhar Mangar
              • ... Yonik Seeley
              • ... Noble Paul നോബിള്‍ नोब्ळ्
              • ... Ryan McKinley
              • ... Noble Paul നോബിള്‍ नोब्ळ्
              • ... Andrzej Bialecki
              • ... Grant Ingersoll
              • ... Ryan McKinley
              • ... Mark Miller
              • ... Noble Paul നോബിള്‍ नोब्ळ्
              • ... Noble Paul നോബിള്‍ नोब्ळ्
              • ... Sami Siren
              • ... Noble Paul നോബിള്‍ नोब्ळ्

Reply via email to