[infinispan-issues] [JBoss JIRA] (ISPN-4650) MassIndexer should not use UpdateDocument when adding to Lucene

[Sanne Grinovero (JIRA)]

    [ https://issues.jboss.org/browse/ISPN-4650?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12993866#comment-12993866 ] 

Sanne Grinovero edited comment on ISPN-4650 at 8/19/14 7:51 AM:
----------------------------------------------------------------

The finding is certainly correct, I've witnesses as well significant slowdowns caused by delete operations.
But for the MassIndexer to ignore put operations which could be triggered "in parallel" has some dangers, you'd need to carefully test this and see if you don't need to queue up update operations which are triggered by normal transactional operations while the MassIndexer is running.

I suspect you'll need to make sure that any update operation triggered by cache write operations needs to be applied after the MassIndexer is done.

For sake of simplicity, you could make this an option which is disabled by default and properly explains the drawback of enabling it: it would be perfectly fine for a batch upload if the user simply knows because of how his application is designed, that no concurrent updates will be applied to the cache while the indexer is running.

Alternatively rather than using a queue - which poses questions on where to store it and how to limit it from uncontrolled size - is to go with the optimistic approach and automatically switch to the safer approach if any other indexing event is generated.


was (Author: sannegrinovero):
The finding is certainly correct, I've witnesses as will significant slowdowns caused by delete operations.
But for the MassIndexer to ignore put operations which could be triggered "in parallel" has some dangers, you'd need to carefully test this and see if you don't need to queue up update operations which are triggered by normal transactional operations while the MassIndexer is running.

I suspect you'll need to make sure that any update operation triggered by cache write operations needs to be applied after the MassIndexer is done.

For sake of simplicity, you could make this an option which is disabled by default and properly explains the drawback of enabling it: it would be perfectly fine for a batch upload if the user simply knows because of how his application is designed, that no concurrent updates will be applied to the cache while the indexer is running.

Alternatively rather than using a queue - which poses questions on where to store it and how to limit it from uncontrolled size - is to go with the optimistic approach and automatically switch to the safer approach if any other indexing event is generated.

> MassIndexer should not use UpdateDocument when adding to Lucene
> ---------------------------------------------------------------
>
>                 Key: ISPN-4650
>                 URL: https://issues.jboss.org/browse/ISPN-4650
>             Project: Infinispan
>          Issue Type: Enhancement
>      Security Level: Public(Everyone can see) 
>          Components: Embedded Querying
>    Affects Versions: 7.0.0.Beta1
>            Reporter: Gustavo Fernandes
>            Assignee: Gustavo Fernandes
>             Fix For: 7.0.0.Beta2
>
>
> The MassIndexer currently issues an Update operation to hibernate search backend, which in turn becomes a delete plus and add in the index.
> Lucene buffers those deletes queries and during merge it tries to 'apply' those deletes wasting a massive amount of time doing seeks and queries unnecessarily.
> Since the mass indexer wipes the index at the beginning, it should simply issue an add operation. Performance wise this make a huge difference: 
> * indexing 50k documents brings down the indexing time from 195s to 33s
> * indexing 200k documents brings down the indexing time from 600s to 55s



--
This message was sent by Atlassian JIRA
(v6.2.6#6264)