[Building Sakai] Recovering search

[Joshua Swink]

Thanks, David. I will be trying this out.

Josh

On Sat, Jan 7, 2012 at 2:12 AM, David Horwitz <david.horwitz at uct.ac.za> wrote:
> Hi Joshua
>
>
> On 01/06/2012 09:46 PM, Joshua Swink wrote:
>>
>> Yes, this is the error I'm seeing. We only have one instance
>> performing indexing and searching, so would it be OK to just shut that
>> one down while deleting the index?
>>
>> 2012-01-04 13:56:05,103  WARN http-8083-Processor24
>> org.sakaiproject.search.journal.impl.JournaledFSIndexStorage -
>> Unlocked Lucene Directory for update, hope this is Ok
>> 2012-01-04 13:56:05,170 ERROR http-8083-Processor24
>> org.sakaiproject.search.component.service.impl.BaseSearchServiceImpl -
>> Failed to get an index searcher
>> org.apache.lucene.store.NoSuchDirectoryException: directory
>> '/local/users/sak/tomcat10c/sakai/indexwork/index-import/38210' does
>> not exist
>>         at
>> org.apache.lucene.store.FSDirectory.listAll(FSDirectory.java:504)
>> ...
>>
> This looks like a corrupt local index
>
> There are 2 properties to configure the search settings:
>
> sharedJournalBase at org.sakaiproject.search.api.JournalSettings  - the shared
> index on multiple noded
> localIndexBase at org.sakaiproject.search.api.JournalSettings - the local index
> for this node
>
> I'd have to check the behaviour if no shared index is specified. The most
> drastic reset you can do is:
>
> 1) Delete the local index
> 2) Delete the share index
> 3) Clear the database entries:
>
> truncate table search_journal;
> truncate table search_node_status;
> truncate table search_segments;
> truncate table search_transaction;
> truncate table searchbuilderitem;
> truncate table searchwriterlock;
>
> (mysql syntax for oracle chage to delete from)
>
> 4) start up
> 5) as admin go to the search admin and select "rebuild whole index)
>
>
>> On Fri, Jan 6, 2012 at 11:39 AM, David Horwitz<david.horwitz at uct.ac.za>
>>  wrote:
>>>
>>> Hi Joshua,
>>>
>>> Do you see any search related errors in the logs?
>>>
>>> One thing you may try is:
>>>
>>> 1) Stop the nodes
>>> 2) delete the local index
>>> 3) Restart the nodes
>>>
>>> This will cause them to reload the index from the shared index
>>>
>>> D
>>>
>>> On 01/06/2012 09:36 PM, Joshua Swink wrote:
>>>>
>>>> Is there a procedure that can be followed to recover the search
>>>> system? The data seems to have gotten into a state where it's
>>>> impossible to search and the system isn't recovering on its own.
>>>>
>>>> Thanks for any assistance.
>>>>
>>>> Josh
>>>> _______________________________________________
>>>> sakai-dev mailing list
>>>> sakai-dev at collab.sakaiproject.org
>>>> http://collab.sakaiproject.org/mailman/listinfo/sakai-dev
>>>>
>>>> TO UNSUBSCRIBE: send email to
>>>> sakai-dev-unsubscribe at collab.sakaiproject.org with a subject of
>>>> "unsubscribe"
>>>
>>> _______________________________________________
>>> sakai-dev mailing list
>>> sakai-dev at collab.sakaiproject.org
>>> http://collab.sakaiproject.org/mailman/listinfo/sakai-dev
>>>
>>> TO UNSUBSCRIBE: send email to
>>> sakai-dev-unsubscribe at collab.sakaiproject.org with a subject of
>>> "unsubscribe"