Redo copy latch, allocation latches & log file sync

What is redo copy latch ?
Generation of redo requires a redo copy latch to be acquired by a process. This is done so that the LGWR knows the data is being copied and so it doesn't flushes that data in the REDO LOG.


What is redo allocation latch ?
Redo allocation latch is obtained to allocated space in the log buffer. This is done to know which log buffer blocks are used and which are free.
If you see this event, it could be due to a number of things.
1) You don't have enough space in your log buffer and processes are waiting for space allocatioin, or lgwr is slow due to slower disks.
2) dbwr gets a latch to see whether these blocks are written to the disk so it can write these blocks to the data files.
3) The sessions that are waiting on log file sync cycle thru and acquire redo allocation latch to check on the log buffer blocks are written to the redo logs or not.

These are the basic reasons why you will see redo allocation latch. Ask me if you need to know how to fix these issues.


What is log file sync ?
Sessions waiting on the return on the commit from the log buffer to guarantee recovery.
DBWR waiting on LGWR to write redo blocks to the redo buffer so dbwr could write the corresponding blocks in the datafiles.
Slower disks
Smaller log buffer.