Join GitHub today
HSEARCH-2146 Waiting until ES indexes have become green #1023
The second commit aims at parallelizing index creation. While it works it feels a bit ad-hoc-ish. Let me know what you think. It cuts down test execution time a bit more, but of obviously the effect will be larger if there are more indexes involved.
Note that I'm handling (index creation + mapping creation) as one execution unit, one could think about splitting these two up, too. This wouldn't improve the end-to-end time, but it'd free threads when waiting for index creation to finish. I don't think it's needed right now.
Mar 8, 2016
@Sanne Reworked some of the things you mentioned and force-pushed.
Note that I've removed the commit which parallelized index creation as I noticed it was causing data races when it comes to indexes shared between different entities. This will need some more consideration.