lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Luis Cappa Banda (JIRA)" <j...@apache.org>
Subject [jira] [Created] (SOLR-4108) SolrCloud: Unexpected behavior when doing atomic updates or document reindexations.
Date Fri, 23 Nov 2012 09:08:57 GMT
Luis Cappa Banda created SOLR-4108:
--------------------------------------

             Summary: SolrCloud: Unexpected behavior when doing atomic updates or document
reindexations.
                 Key: SOLR-4108
                 URL: https://issues.apache.org/jira/browse/SOLR-4108
             Project: Solr
          Issue Type: Bug
          Components: SolrCloud
    Affects Versions: 4.0
         Environment: Note: This issue is related with JIRA-4080.
Context: SolrCloud deployed with nShards=1, two Solr servers, each one with two cores/collections.
We haven then one leader and one replica for each shard.
            Reporter: Luis Cappa Banda
            Priority: Blocker
             Fix For: 4.1


The situation is this the following:

1. SolrCloud with one shard and two Solr instances.
2. Indexation via SolrJ with CloudServer and a custom BinaryLBHttpSolrServer that uses BinaryRequestWriter
to execute correctly atomic updates. Check JIRA-4080
3. An asynchronous proccess updates partially some document fields. After that operation I
automatically execute a commit, so the index must be reloaded.

What I have checked is that both using atomic updates or complete document reindexations random
documents are not updated even if I saw debugging how the add() and commit() operations were
executed correctly and without errors. In other words, something strange happens when you
both index and update documents asynchronously at the same time.

Also, if I debug line by line and I check with my own eyes when an index operation is done,
I confirm that the document itself updates correctly.

What I think is that there is some critical problem with both SolrCloud and CloudSolrServer
interface that has something to do with index blocking while writing and forwarding document
updates to replicas.

If I´m right, and considering also JIRA-4080, I would not recommend SolrCloud in production
at the moment.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org


Mime
View raw message