Rebuilding Solr indexes

Eric Redmond eredmond at basho.com
Tue Jan 27 17:02:11 EST 2015


> On Jan 27, 2015, at 3:21 AM, Sorin Manole <sorin.manole at trustev.com> wrote:
> 
> Hi all,
> 
> We are using Solr indexes on top of Riak keys to make the search more easier.

Just to clarify, are you using Riak Search (aka, Yokozuna), or are you managing Solr separately? I’m assuming the former.

> One of the doubts regarding Solr is rebuilding the indexes. That could happen when we want to add a new index to the schema - which would mean drop and recreate the indexes. 

Do you mean adding a new field to the schema? You have a few options after updating the schemas.

1) Create a new index and reattach, as you suggested
2) Rewrite Riak values inline, which will cause them to re-index
3) Delete the index directories on every node and let AAE rebuild them

There’s some new code that will create a 4th option https://github.com/basho/yokozuna/pull/427 <https://github.com/basho/yokozuna/pull/427>

4) update the schema and let AAE fix the index

Then, we're working on a 5th option:

5) modify schemas via command-line, and start an inline, eventually consistent index update

There is an existing discussion on this topic:

http://comments.gmane.org/gmane.comp.db.riak.user/15420 <http://comments.gmane.org/gmane.comp.db.riak.user/15420>


> As far as we know after dropping and recreating the Solr indexes, for the already existing keys in Riak thos indexes don't rebuild. Is there a way to rebuild indexes on already existing keys when recreating a Solr index?

See above.

> Also, if that's possible, any idea how time costly that'd be? Let's take example rebuilding indexes for 100 million riak keys.

This is hard to say. Depending on the number of nodes, hardware, bandwidth, object size, etc. If they’re small objects, it could be a few minutes.

> Another doubt is around corrupting Solr indexes? How "probable" is it? Same question as before.. in the case of a corrupted index, how easy would be to re-index 100 million keys? 

You shouldn’t need re-index anything. If values in an index are corrupted, the AAE (active anti-entropy) service will repair the values automatically.

> Thanks in advance.
> 
> -- 
> 	
> Sorin Manole
> Senior Software Engineer, Trustev
> m:+353 86 051 2658 <> | e:sorin.manole at trustev.com <mailto:sorin.manole at trustev.com> | w:www.trustev.com <http://webapp.wisestamp.com/www.trustev.com>| a: Trustev Ltd, 2100 Airport Business Park, Cork, Ireland.
> 
> This message is for the named person's use only. If you received this message in error, please immediately delete it and all copies and notify the sender. You must not, directly or indirectly, use, disclose, distribute, print, or copy any part of this message if you are not the intended recipient. Any views expressed in this message are those of the individual sender and not Trustev Ltd. Trustev is registered in Ireland No. 516425 and trades from 2100 Cork Airport Business Park, Cork, Ireland.
> 
> _______________________________________________
> riak-users mailing list
> riak-users at lists.basho.com
> http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.basho.com/pipermail/riak-users_lists.basho.com/attachments/20150127/969a9ae9/attachment-0002.html>


More information about the riak-users mailing list