We have an oracle 126.96.36.199 server with 8 cores and 12gb ram. It's CentOS 5.6 and it has 12gb swap file. Oracle's memory target is 4gb and memory management is automatic. It has MegaSAS Raid controller with 4 drives.
Database itself is around 15gb with lots of blob data.
During sql selects server hangs a lot with high waiting for io, around 100%. It hangs in a way that I can't even connect to it using ssh. It eats around 5-6 gigs of swap. selects doesn't include blob data only plain rows.
vm.swappiness = 0 and I even tried
swapoff but it still hangs.
What might be the problem?