Yeah, it's strange because we definitely have periods of high iowait but this is not when the locks are happening. If I could correlate it directly to that it would be so much easier. Thanks again for the response!
On Wed, Feb 1, 2012 at 8:42 PM, Alan Hodgson
<ahodgson@simkin.ca> wrote:
On Wednesday, February 01, 2012 05:13:15 PM Christopher Opena wrote:
> Do you mean 6-12% of total iowait, or per cpu? Our average iowait in the
> last week is 34.31% of a total 1600% with an average idle of 1451.76%. Our
> iowait *does* spike occasionally (today it went up to 148.01%) but it
> doesn't coincide with the lock happening. At the time of the lock we were
> at 10.58% iowait, which is quite a bit below our average.
>
Total, but it doesn't sound like that's the problem.