On Tue, Oct 27, 2009 at 4:54 PM, Tomas Bodzar <bodzart at openbsd.cz> wrote:
> Maybe it will be better to focus on prstat because of this?
>
> http://www.brendangregg.com/DTrace/prstatvstop.html

My own experience runs counter to that. I'm not saying that the
analysis is incorrect (although I've done several such analyses
and sometimes top wins, sometime it loses, so it's nowhere near
as clear cut).

What I find is that top is more capable and useful. To get the same
level of information out of prstat is much more expensive than
running top - you either have to run multiple copies of prstat, or
keep quiting and restarting to get a different view, or run additional
tools as well, which far outweigh any savings which may or may not
be present due to implementation details.

(Although I have recently noticed that prstat on my production boxes
has been chewing resources at a rate significantly in excess of top.
The biggest danger with top is that it's such a useful tool that you're
tempted to leave it running permanently.)

-- 
-Peter Tribble
http://www.petertribble.co.uk/ - http://ptribble.blogspot.com/

Reply via email to