Siebel on Oracle DB - what are normal values for "DB File sequential read" wait events
We are using Siebel 8.1.1.6 on ORACLE 11.2.0.2 (Windows Server 2008 R2). There are ~ 150-200 concurrent connected users. The SBL_DATA and SBL_INDEX Tablespace are approx. 80GB in total. The server has 64GB RAM from which 50GB is assigned to be managed by AMM.
Now sometimes, queries take a long time to respond. Long time means > 15 seconds which is kind of too long for a simple query in my opinion. Analysing the execution plans, there are no full table scans, just regular index access. Almost all time is spent with DB file sequential read wait event. The DB server is connected to a HP EVA storage which *should* be quite fast, but obviously it isn't? I ran some AWR reports and User IO Wait (especially "DB file sequential read") is always the top wait event recorded. It takes from 30% up to 75% of total