Craig Shallahamer's Blog
|
You were brought to this page based on an internet search
and as a free service to Oracle DBAs.
The text below is an except from the book,
Oracle Performance Firefighting, written by
Craig Shallahamer of
OraPub, Inc.
Figures and tables are not included on this page, only their reference.
To order the book in either print or PDF form, click
here.
©2009, 2010 by Craig Shallahamer. This is copyrighted material.
Please—Out of respect for those involved in the creation of the book and also for
their familes, we ask you to respect the copyright both in intent and deed. Thank you.
-------------------------------
<p>If a process is spending too much time sleeping while acquiring a redo allocation latch, the latch free wait event associated with the redo allocation latch (pre-Oracle Database 10g) or the latch free: redo allocation (Oracle Database 10g and later) wait event will be your top wait event. This is most likely to occur in an intense high-concurrency environment. Fortunately, there is a straightforward Oracle-focused solution, which works with any Oracle release, that will probably resolve the issue.
p><p>The bad news is, as discussed earlier, pre-Oracle9i Release 2 systems have a single redo allocation latch, making it much more likely to experience redo allocation latch contention. The good news is there are multiple redo copy latches, and by creatively adjusting the _log_small_entry_max_size parameter, we can influence the likelihood of a process having to acquire the redo allocation latch.
p><p>Think of this instance parameter as a fence. On one side of the fence, the allocation latch is used, as you would expect, for allocating space in the redo log buffer, and one of the redo copy latches is used while copying redo into the just-allocated space. On the other side of the fence is the redo copy latch, which is used to both allocate redo log buffer space and to copy redo into the just-allocated space! If the redo entry size is greater than the _log_small_entry_max_size value, the process will hop over the fence, using only one of the redo copy latches and thereby bypassing requesting the redo allocation latch. So, by reducing the height of the fence-that is, reducing the parameter value-the processes are less likely to request a redo allocation latch. This directly reduces redo allocation latch concurrency, which will reduce the likelihood of receiving the latch free wait event.
p>
©2009, 2010 by Craig Shallahamer. This is copyrighted material.
Please—Out of respect for those involved in the creation of the book and also for
their familes, we ask you to respect the copyright both in intent and deed. Thank you.
|
|
Know what's important before it's too late!
OraPub's
Performance Training
is like no other...
More Class Pics...
Get student testimonials!
|
