[Cialug] How to determine the cause of a Linux performanceproblem?

[Pixel // pinterface]

Paul Gray wrote:
> On 02/27/2012 07:49 AM, Pixel // pinterface wrote:
>
>> It appears (via top) that the processes bumping up the load average
>> by being in wait states are often "kjournald" and "flush-###",
>>
>
> The above would imply a lot of logging activity.  It could either be
> the direct cause due to frantic logging, or some system event that
> would be inducing a lot of logs.
>
> Are any of your logs "exceptionally" large in comparison to previous
> log files?

Alas, no.  The logs all look pretty normal in both size and content.

Too bad.  Something being excessively logged probably would have been an
easy fix. :)

--