Hi,
We have written a multi-threaded indexing framework that uses Lucene. To avoid contention on a single IndexWriter, we are using multiple writers. We have enountered a situation where we need to check if a writer object exists for a particular index directory and create it if it doesn't. We do this by catching the exception that IndexWriter throws. We found from the code that the WRITE_LOCK_TIMEOUT is one second. This value is a bit high for our requirements.
My query is
* Why is it hard-coded to minimum 1 sec.
* Will anything break if I change to value to be very low, say 5 millisecond and use it.
Thanks
~ Mukul Joshi
Senior Member of Technical Staff
Great Software Laboratory, Pune
We have written a multi-threaded indexing framework that uses Lucene. To avoid contention on a single IndexWriter, we are using multiple writers. We have enountered a situation where we need to check if a writer object exists for a particular index directory and create it if it doesn't. We do this by catching the exception that IndexWriter throws. We found from the code that the WRITE_LOCK_TIMEOUT is one second. This value is a bit high for our requirements.
My query is
* Why is it hard-coded to minimum 1 sec.
* Will anything break if I change to value to be very low, say 5 millisecond and use it.
Thanks
~ Mukul Joshi
Senior Member of Technical Staff
Great Software Laboratory, Pune