solr Indexing taking too much time...

[arpit_gupta]

Hello friends,

I have integrated REST services with solr subsysten in my application.

I am uploading documents with metadata  using default upload and metadata scripts. and after uploading data and updating metadata i search alfresco for documents on the basis of metadata.

Now the problem is alfresco is not returning currently uploaded document in search results and hence i can not show the correct document list on application UI and that's why user think that document is not uploaded.

Though if i fire the search script 2-3 times more i get the correct result. I guess alfresco is taking time in indexing the metadata.

I am in real trouble here. Any help??

Thanks,
Arpit

[mrogers]

If you are using SOLR you must take account of the fact that there will be a delay (however short) between uploading content and being able to search for it. 

There are a couple of approaches.  
a) Don't use search to find a document you just uploaded, use one of the other APIs such as a lookup by noderef or list the contents of a folder.    That's how Alfresco Share works.
b) Don't use SOLR, use the in transaction Lucene index subsystem instead.

[lesignor]

Can you help me to find a javascript api to lookup for a new node from its noderef, something like search.findNode(..), but this API uses lucene index to work.

Thank you

[arpit_gupta]

i have to do search on metadata and i have to show document list on the basis of that metadata.
Hence a search on folder or noderef will not work.

So that means i can not use solr subsystem in this scenario 
Isn't there any other way to resolve this problem without dropping solr?

[mrogers]

No.

[acarpine]

mroger really thanks for your clarifications.
What is not clear for me is how to specify/change how often solr must run.

From http://wiki.alfresco.com/wiki/Alfresco_And_SOLR i thought that the property was

alfresco.cron=0/15 * * * * ? *
    How often to check if there is work to do

Although alfreco.cron sets the interval every 15 minutes, in my Alfresco installation I'm able typically to retrieve a node by metadata just after 20/30 seconds while in other installations the time to wait seems significantly longer. Could you explain what's the correct parameter or where this parameter is overridden?

Thanks,
Andrea

[mrogers]

I think that cron expression is 15 seconds.

[acarpine]

Ops right…Many thanks mrogers!

[srm]

I am in the same boat as some one, where an uploaded document is not immediately available for search. I am using the SOLR subsystem as well. As i see from earlier answers, i should consider changing the subsystem from solr to lucene. Could anybody tell, how to change the subsystem from solr to lucene? are there any disadvantages by doing this?  I am using Alfresco 4.0 version.

Is it true that indexing is less important in Alfresco 4.0 and it is advised to turn off this feature?

[mrogers]

Could anybody tell, how to change the subsystem from solr to lucene?
    Please see the documentation for the search subsystem.
-> are there any disadvantages by doing this?
    Performance is affected by having in transaction indexing,  you can't index offline or share indexes between a cluster of servers.
Is it true that indexing is less important in Alfresco 4.0 and it is advised to turn off this feature?
-> Its not essential like it used to be but search is an important part of alfresco.