Closed adf-ncgr closed 6 years ago
The Lucene indexes were built. Lucene seems to be broken.
I don't think there's much I can do on this, other than maybe discover some stuff that's being indexed that shouldn't be. It takes forever on dev, and that's faster than production. We'll keep pondering.
Well, here's one reason, probably amongst others:
20-Dec-2017 15:57:32.845 SEVERE [ajp-apr-8009-exec-232] org.apache.catalina.core.StandardWrapperValve.invoke Servlet.service() for servlet [action] in context with path [/legumemine] threw exception [javax.servlet.ServletException: java.lang.OutOfMemoryError: Java heap space] with root cause
java.lang.OutOfMemoryError: Java heap space
I'll boost the RAM on shokin-webapps.
So this seems to be solved. I had somehow neglected to omit Homologue and ExpressionValue from Lucene indexing, which made the indexing take forever and the search take forever. Blocking those two classes, plus a bunch of smaller ones, brought indexing down to 15m on production and searches seem to respond pretty quickly now.
It was mainly Homologue bloat from adding all the new organisms, with a touch of expression bloat from the additional expression sets.
much improved! no more naptime for @jd-campbell :)
whenever is most appropriate. I won't be relying on them for what I'm doing