yes, would be much better.

Jokin

On 7/2/07, Oliver Castle <[EMAIL PROTECTED]> wrote:
Hello

Sorry about being so vague, currently we open a index searcher every
time we want to query the index which I am guessing maybe the cause of
the problem after reading some earlier threads. Would we be best to
create something like a singleton index searcher and then have all
threads access the one reader?

Cheers

Ollie Castle

Jokin Cuadrado wrote:
> You are reading or writing the index? could yo be a bit more specific
> about your configuration? (when you open a index reader or writer, if
> you use the same reader to delete documents and search them, you
> access with different readers to the index, etc.)
>
> We have tested lucene.net in high load environments during almost a
> week and never had any lock problems, even so, one of the guidelines
> to obtain the better performance from lucene.net is to share the same
> indexreader over all the reading threads, so lucene can maintain
> caches of more accessed items and don't need to read from disk.
>
> --
> jokin
>
>
>
>
>
>
> On 7/2/07, Oliver Castle <[EMAIL PROTECTED]> wrote:
>> Hello
>>
>> We are currently developing a Lucene.Net search for one of our products
>> and we are very happy with results apart from when we we load test the
>> application at high loads and the application starts the throw errors
>> returning that the lucene index still has a lock from another IIS thread
>> on it. Do we have to implement locking between IIS threads so only one
>> IIS request at a time can read the index or are we doing something wrong
>> in the code?
>>
>> Any suggestions great fully received.
>>
>> Ollie Castle
>>
>>
>



Reply via email to