Cannot run certain queries after upgrade

What steps will reproduce the problem?
I'm not 100% sure how to reproduce this problem in an environment other than my 
own, but a few queries that produce the error are:
eventid:4625
eventid:1100 eventid:1108
(Interestingly enough, 'eventid:1100 OR eventid:1108' produces no error)
eventid:4722
eventid:4697
WINDOWS.eventid=4722 (same for all the other single eventid queries)
class=BRO_FILE

What is the expected output? What do you see instead?
All of the errors produce something like this:
Errors: 127.0.0.1 had 39 indexes had a non-existent fields: attr_i0
Or, for the 'class=BRO_FILE' query:
Errors: 127.0.0.1 had 41 indexes had a non-existent fields: class

The number of indexes with the non-existent fields has been going up steadily.

What version of the product are you using? On what operating system?
r950 with two nodes. 127.0.0.1 is running the latest CentOS.

Please provide any additional information below.
I did also recently upgrade mysql to MariaDB, but I don't think that's the 
issue. I can try rolling back if you need me to, though.

Original issue reported on code.google.com by i...@pingas.org on 2 Jul 2013 at 1:55

MHMDhub / enterprise-log-search-and-archive

Cannot run certain queries after upgrade #162