Hello,
We’re running Infobright with default settings for a 16G server (10000, 1000, 800).
Our queries are always like this
INSERT INTO memorytable{$i} SELECT commonfieldID, field0 FROM table{$i}_1 INNER JOIN table{$i}_2 ON commonfieldID WHERE field1 AND field2 AND field3 AND (field4 OR field5 OR field6)
- {$i} is 1 to 8
- table{$i}_1 and table{$i}_2 have the same number of rows and joined on a unique id
- the tables have about 3 million rows each.
We run all the 8 queries at the same time.
The problem is that for high number of hits (>1 million) for field4, field5, field6 (the ones in the group with OR operator) the server crashes.
I’ve checked bh.err and the memory size that the engine tries to allocate is available on the server. We’ve tried adjusting the settings in /etc/my-ib.cnf but with no success.
Can you tell me if we’re doing something wrong here?
Thanks,
Justin

