[dundee] Investigating resource hungry processes

[Lee Hughes]

I've used this 

http://www.linux.com/archive/feature/136114

it got results..also http://oprofile.sourceforge.net/examples/
there's also a few more kernel profilers, that may shed some light
on the situation if your bottle necks are in kernel space rather than user space.


--- On Mon, 27/7/09, Andrew Clayton <andrew at digital-domain.net> wrote:

From: Andrew Clayton <andrew at digital-domain.net>
Subject: Re: [dundee] Investigating resource hungry processes
To: dundee at lists.lug.org.uk
Date: Monday, 27 July, 2009, 7:28 PM

On Mon, 27 Jul 2009 19:06:43 +0100 (BST), Marcel Hecko wrote:

> Hi Iain,
> 
> we are using `iotop`, `ps` (ps -efH) and `lsof` (lsof -i for sockets)
> to digg into the resources usage on the system.

Yeah, all good.  strace is also useful. Like others, top is about the
very thing I look at, gives a good overall look at the current
situation.

Then of course you have /proc/PID with an abundance of information
(where most of the above get their information from).

Your lsof is useful in conjunction with /prot/PID/fd (list of open file
descriptors for that process)

pmap can be useful to view a processes memory mapping.

I highly recommend to install the sysstat package, which will do
periodic system stats gathering which can be queried with the sar
command (people from a Solaris background will probably recognise this).

This also includes iostat (like vmstat, but for blockdevices/filesystems).


Andrew

_______________________________________________
dundee GNU/Linux Users Group mailing list
dundee at lists.lug.org.uk  http://dundeelug.org.uk
https://mailman.lug.org.uk/mailman/listinfo/dundee
Chat on IRC, #tlug on irc.lug.org.uk



      
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.lug.org.uk/pipermail/dundee/attachments/20090727/b34ff49c/attachment.htm