urgent issue - redo buffer allocation retries 1

[orauxpro]

hi

we are installing hp-ux9000 v.11, oracle 8.0.04, i have a problem with the retries counter, i have 4 redo log files, and each 2m size, i have 160 users accessing the database, the value if retires keeps up can anybody tell me what area in oracle should i look in, Please help me as soon as posible.

Oraux

 

[karluk]

Your redo logs seem small to me. How often are you doing log switches? Log switches are recorded in the alert log, so you can check there. If you are switching every minute or two during peak times, then I would say that your redo logs are too small. I would increase them to, say, 20M.

The redo buffer is actually different from the redo log files. It is an area of memory (in Oracle's SGA) set aside to hold redo log info. The size is controlled by the initialization parameter "log_buffer". Problems with allocation retries suggest that this buffer may be too small. However, I would first deal with the possibility of excess log switches. Waits for log switches may be showing up as problems with log buffer allocations.