high "Load Average"

[Sessna]

Stefan Parvu Wrote:
-------------------------------------------------------
> You should check CPU%, Disk IO% and probable look for 
> operations on the wait queue. sysrec does this on Solaris, 
> Im working to port the recorder on Linux, FreeBSD.
> You should somehow count for yourself:

I am using sar tool at the moment and unfortunately not that experienced to migrate SDR from Solaris myself

Here is snippet from sar output for typical situation when the system slows down:

CPU     %user     %nice   %system   %iowait    %steal     %idle
all      7.92     0.00    7.75      46.93      0.00       37.40
all      8.43     0.00    8.31      48.12      0.00       35.14
all      8.03     0.00    8.01      45.84      0.00       38.12
all      8.39     0.00    8.43      46.48      0.00       36.69
all      9.89     0.11    7.32      54.83      0.00       27.85
all      9.09     0.00    8.87      49.75      0.00       32.30
all      8.96     0.00    9.09      48.89      0.00       33.06
all     10.02     0.00    9.60      51.23      0.00       29.15



>  - Disk IO, CPU% and look for signs of saturation.
>  - cpuplayer (is your machine spending more time in SYS/USER ?)
>  - useful to collect longer trends in order to see whats going on

>From this table we can tell that CPU is not completely used, there is about 30% idle, user and system CPU consumptions are almost equal (is it good or bad? what usage ratio is expected?)  %iowait looks like high (relatively to other metrics) but I am not quite sure what does it mean. 
Does anybody see something leading to slowness in these metrics?

Thanks,
Sessna

Posted at Nginx Forum: http://forum.nginx.org/read.php?2,63176,63668#msg-63668