Oracle redo allocation latch misses Oracle Database Tips by Donald Burleson

 

Question: 

 

In my STATSPACK report I see high miss values for redo allocation latch.  What do I do to relieve the misses on the redo allocation latch?

 

Answer:

 

In the latch section of a STRATSPACK or AWR report you should look for the redo allocation line where ?pct get miss? > 0.1 OR ?pct nowait miss? > 1.0.

 

The redo allocation latch controls the allocation of space for redo entries in the redo log buffer (as defined by the log_buffer parameter).  The redo allocation latch is a serialization latch that enforces the sequence of entries in the log buffer.  A process can only write a redo entry after the redo allocation latch, which is why it's a will-to-wait latch.

 

Redo allocation latches are a willing-to-wait latch, and the traditional remedy was to increase the size of the log_small_entry_max_size parameter, but this parameter was removed in Oracle8i.

 

You can measure redo allocation latch misses with this query:

 

select
round(
   greatest(
     (sum(decode(ln.name,'redo allocation',misses,0))
/greatest(sum(decode(ln.name,'redo allocation',gets,0)),1)),
(sum(decode(ln.name,'redo allocation',immediate_misses,0))
/greatest(sum(decode(ln.name,'redo allocation',immediate_gets,0))
+sum(decode(ln.name,'redo allocation',immediate_misses,0)),1))
)*100,2)
from
   v$latch l,
   v$latchname ln
where
   l.latch#=ln.latch#;

 

Oracle consultant Steve Adams notes on on redo allocation latch tuning:

"The redo latches are not only taken in connection with redo generation and log file sync waits. They are taken by LGWR as well in connection with writing redo from the log buffer to the log files.

When LGWR wakes up, it first takes the redo writing latch to update the SGA variable that shows whether it is active. This prevents other processes from posting LGWR needlessly. Then, if it was not posted, LGWR then takes the redo allocation latch to determine whether there is any redo to write. If not, it takes the redo writing latch again to record that it is no longer active, before starting another rdbms ipc message wait.

If there is any redo to write, LGWR then inspects the latch recovery areas for the redo copy latches (without taking the latches) to determine whether there are any incomplete copies into the redo buffers that it intends to write. If so, LGWR sleeps on a LGWR wait for redo copy wait event, and is posted when the required copy latches have been released. The time taken by LGWR to take the redo writing latch, the redo allocation latch and to wait for the redo copy latches is accumulated in the redo writer latching time statistic."