we are running OPS in 126.96.36.199.0 on a Compaq Tru64 5.0A.
The redo log buffers are sized at 163840.
The 'redo buffer allocation retries' seem to be on a high side. The query
SQL> select inst_id,value from gv$sysstat where name='redo buffer allocation retries';
returns the following:
Is it necessacry, that the log_buffers should be increased from the present value?
How does retaining the same will affect the performance?
Oracle reccommends that retries should be less than 1% of redo entries. If is not then it should be OK to increase the log_ buffer. If it is then increasing the log_buffer will not result in any noticeable benefits. I think you need to collect a bit more statistics.
How can I know the redo entries?
And can you provide links to the doc. where it says it should be 1% of the entries? Because, I could only get "The value of REDO BUFFER ALLOCATION RETRIES should be near zero. "