I wrote a small program to index a 100k document with
name:foo bar" + TestUtil.randomSimpleString(random, 5, 10)
This was the query I tried performing :
q={!surround maxBasicQueries=100000 cache=false} w(foo, bar*)
I tried it on branch_5x and trunk and found very different results
*branch_5x* - the query would take about a minute and then hit an OOM with
a 512M heap .
*trunk *- query runs within 3/5 seconds consistently without hitting OOM's.
It roughly takes around 120MB to execute the queries. Additionally I tried
to move the branch_5x index to trunk and then run the same query and got
the similar fast results as well.
Here is a VisualVM screenshot of both . The top image being of trunk and
the bottom image being branch_5x : http://imgur.com/a/iKxmL
Any pointers as to what could be the reason for this vast improvement?
--
Regards,
Varun Thacker
