Gather stats concurrent program scheduled on weekly basis?
if not run the program and check the performance
Yes, it is running
Describe your architecture please.
Are you using a separate middle tier from the database tier?
Are you using Linux? HugePages/Large Pages on other Operating Systems?
If on Linux, please run free -m and report memory usage next time before you restart services and after.
Database is in one node and application in other server.
If memory (free -m) high, such case what need to check?
You need to look at -/+ buffers/cache line. In this case, the server had 256G and about 108G is used, so about 150G free.
total used free shared buffers cached
Mem: 258445 257492 953 0 145 149149
-/+ buffers/cache: 108196 150248
Swap: 16385 748 15637
As mentioned, Gather Stats & CPU/RAM utilization at OS level to be checked for any resource crunch.
1. Please go through AWR report (for 60 min interval) during the problematic period. This is to check any expensive DML/Query in the system (or) Concurrent Manger sub-system.
2. Please go through the output of bde_chk_cbo.sql as per note 174605.1 to check whether init.ora are in tact.